Chapter 3

Review of the couesi Complex (Rodentia: : ) in Western Mexico

MICHAEL D. CARLETON1 AND JOAQUIN ARROYO-CABRALES2

ABSTRACT

The status and distribution of eight -group taxa of the group (Rodentia: Cricetidae: Sigmodontinae) described from western Mexico are evaluated based on morphological and morphometric comparisons. Four of these are recognized as valid species within the region: Merriam, 1901 (including molestus Elliot, 1903), from inland plateau of the Mesa de Ana´huac; O. couesi mexicanus J.A. Allen, 1897 (including bulleri J.A. Allen, 1897; lambi Burt, 1934; rufus Merriam, 1901), from Pacific coastal plain, contiguous lower mountain slopes, and interior valleys along the Rı´os Tepalcatepec–Balsas; O. nelsoni Merriam, 1898, from Isla Marı´a Madre; and O. peninsulae Thomas, 1897, from the southern tip of the Baja California Peninsula. Three other taxa named from uplands in interior Mexico— aztecus Merriam, 1901; crinitus Merriam; 1901; and regillus Goldman, 1915—are provisionally retained within O. couesi, but further study of their specific stature and relationships is required. The recommended taxonomic changes within western Mexico serve to discuss directions for future revisionary research that will refine the definition and distribution of O. couesi sensu stricto in Middle America.

INTRODUCTION the nominotypical species and six geograph- ically isolated specific allies within Middle North and Middle American populations America (i.e., O. antillarum, O. cozulmelae, of the Oryzomys palustris group (sensu O. fulgens, O. gatunensis, O. nelsoni, and O. Merriam, 1901; Goldman, 1918) have been peninsulae). Goldman’s specific classification allocated to as many as 18 species (Merriam, was essentially maintained through the mid- 1901) or to as few as three (Hall, 1981). The dle 1900s (Miller and Kellogg, 1955; Hall and systematic contributions of Goldman (1918) Kelson, 1959). Subsequently, Hall (1960) and Hall (1960) have played historically influential roles in reversing the early view considered couesi to be synonymous with O. of specific diversity within the O. palustris palustris based on perceived intergradation of group. In his revision of the , Goldman populations along coastal in certain (1918), perhaps emulating the revisionary cranial and chromatic traits. Although ar- model of as set forth by his gued within a very localized geographic colleague Osgood (1909), demoted many of context, Hall swept all of Goldman’s (1918) Merriam’s species to intergrading geographic of couesi, including those in races of the widely ranging, polytypic forms western Mexico and Central America, under O. couesi and O. palustris. Within the O. his expansive definition of O. palustris. His palustris group, Goldman designated two action set a broad-brushed precedent for sections, the O. palustris section, which assessing and taxonomically conveying mor- contained the single North American species, phological variation among populations and the O. couesi section, which contained of the Oryzomys palustris group. Soon

1 Department of Vertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560- 0111 ([email protected]). 2 Laboratorio de Arqueozoologı´a ‘‘M. en C. Ticul A´lvarez Solo´rzano,’’ Subdireccio´n de Laboratorios y Apoyo Acade´mico, INAH, Moneda # 16, Col. Centro, 06060 Me´xico, D. F., Me´xico ([email protected]).

94 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 95 afterward, other Middle American forms addition, we discuss possible systematic af- retained as species by Goldman (1918) or finities of other taxa named from uplands in Hall and Kelson (1959) were pro forma interior Mexico—O. c. regillus (Michoaca´n), brought into O. palustris as well (cozumelae O. c. aztecus (), and O. c. crinitus by Jones and Lawlor 1965; azuerensis and (Distrito Federal). gatunensis by Handley, 1966; fulgens, nelsoni, and peninsulae by Hershkovitz, 1971; antil- larum by Hall, 1981). MATERIALS AND METHODS Judged against today’s standards of sys- Specimens that form the basis of this tematic documentation, presentation of data report principally consist of skins with their and analysis to support these sweeping associated and are contained in the taxonomic changes has been at best perfunc- following institutional and museum collec- tory, at worst nonexistent. Reexaminations tions (abbreviation in parentheses): American of the purported zone of intergradation Museum of Natural History, New York between couesi (aquaticus) and O. palustris (AMNH); Carnegie Museum of Natural (texensis) in Texas tellingly exposed the History, Pittsburgh (CM); Dickey Collection, fragile underpinnings of such an inclusive University of California, Los Angeles definition of O. palustris (Benson and Gehl- (UCLA); Museum of Natural History, Uni- bach, 1979; Schmidt and Engstrom, 1994). versity of , Lawrence (KU); Field As a result of these regional studies, O. couesi Museum of Natural History, Chicago (Alston) was resurrected as a species distinct (FMNH); National Museum of Natural from North American O. palustris (Harlan), History, Smithsonian Institution, Washing- but most other Middle American taxa have ton, D.C. (USNM, formerly the U.S. Na- remained as synonyms of a broadly ranging O. couesi (e.g., Honacki et al., 1982; Corbet tional Museum); the Natural History Muse- and Hill, 1991; Musser and Carleton, 1993, um, London (BMNH, formerly the British 2005), an arrangement based on classificatory Museum of Natural History), and the Uni- inertia rather than on actual documentation versity of Michigan Museum of Zoology, of consanguinity. Ann Arbor (UMMZ). Type specimens and Herein, we review morphological and original series of most taxa described for morphometric evidence for intergradation populations of Oryzomys in western Mexico among forms of the Oryzomys couesi complex were examined firsthand, namely those of whose distribution lies in western Mexico. albiventer Merriam, 1901; aztecus Merriam, The region encloses striking topographic and 1901; bulleri Allen, 1897; crinitus Merriam, biotic diversity and accordingly has proven to 1901; mexicanus Allen, 1897; nelsoni Mer- be strategic for unraveling specific diversity in riam, 1898; peninsulae Thomas, 1897; regillus other taxonomically problematic gen- Goldman, 1915; and rufus Merriam, 1901. era, such as Peromyscus (Carleton, 1977; Original topotypes of lambi Burt, 1934, were Carleton et al., 1982; Bradley et al., 1989, consulted, but we did not personally examine 1996) and Sigmodon (Zimmerman, 1970; the holotype; William T. Stanley examined Carleton et al., 1999; Peppers et al., 2002). and measured the FMNH type of molestus Eight type specimens of species-group epi- Elliot, 1903, for us. thets currently allocated to Goldman’s (1918) Approximately 600 specimens were exam- O. couesi section originate from this area (see ined (see registration numbers and full fig. 1). In particular, we address the status locality information under Taxonomic Sum- and relationships of taxonomically recog- mary) and measured, of which 284 adults nized populations of Oryzomys that occur at were grouped into 11 geographically cohesive the southernmost tip of Baja California (O. samples for univariate and multivariate couesi peninsulae), on the Tres Marı´as Islands analyses. According to current taxonomic (O. nelsoni), along coastal lowlands in understanding, these operational taxonomic (O. c. lambi) and from through units (OTUs), their abbreviations as em- (O. c. mexicanus), and on the inland ployed in figures, specific localities, and plateau region of (O. c. albiventer). In sample sizes are recognized as follows: 96 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Fig. 1. Type localities of 11 taxa of the Oryzomys couesi species group described from western and central Mexico and discussed in the present study. Taxa and localities include: albiventer Merriam (1901) from Ameca, Jalisco; aztecus Merriam (1901) from Yautepec, Morelos; bulleri Allen (1897) from Valle de Banderas, ; crinitus Merriam (1901) from Tlalpan, Distrito Federal; lambi Burt (1934) from San Jose´ de Guaymas, Sonora; mexicanus Allen (1897) from Tonila, Jalisco; molestus Elliot (1903) from Ocotla´n, Jalisco; nelsoni Merriam (1898) from Isla Marı´a Madre, Nayarit; peninsulae Thomas (1897) from Santa Anita, Baja California; regillus Goldman (1915) from Los Reyes, Michoaca´n; rufus Merriam (1901) from Santiago, Nayarit. Topography of Mexican sierras is indicated by 1000 m elevational bands.

Oryzomys couesi albiventer: (J3) Jalisco, Oryzomys couesi peninsulae: (BC) Baja Ameca, N 5 12. California Sur, San Jose´ del Cabo, N 5 14. Oryzomys couesi mexicanus: (C1) Colima, Oryzomys nelsoni: (TM) Nayarit, Tres vicinity of Santiago, N 5 33; (C2) Colima, 5 Marı´as Islands, Isla Marı´a Madre, N 5 4. mi NW Manzanillo, N 5 20; (J1) Jalisco, Four external dimensions were transcribed localities to the N and E of Barra de from skin tags to the nearest whole millimeter Navidad, N 5 60; (J2) Jalisco, Cuitzamala, (mm): total length (TOTL); tail length (TL); N 5 20; (N1) Nayarit, 1 mi S Cuautla, N 5 hindfoot length (HFL); and ear (pinna) 62; (N2) Nayarit, San Blas and vicinity, N 5 length (EL). 17 cranial and three dental 23; (N3) Nayarit, Chacala and Las Varas, N variables were measured to 0.01 mm (most 5 16; (N4) Nayarit, 2 mi E San Pedro are defined and illustrated in Carleton and Lagunillas, N 5 20. Musser, 1995, and Musser et al., 1998), using 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 97 handheld digital calipers while viewing crania unknowns in the discriminant-function anal- under a stereomicroscope. These measure- ysis and their probabilities of group mem- ments, and their abbreviations as used in text bership calculated a posteriori, as were the and tables, are: occipitonasal length (ONL); holotypes of aztecus, bulleri, crinitus, molestus greatest zygomatic breadth (ZB); breadth of (as measured by W.T. Stanley, FMNH), braincase (BBC); depth of braincase (DBC); regillus, and rufus. Principal components breadth across occipital condyles (BOC); were extracted from the variance-covariance interorbital breadth (IOB); length of rostrum matrix, and variable loadings are expressed (LR); breadth of rostrum (BR); depth of as Pearson product-moment correlation co- rostrum (DR); length of efficients of the extracted components or (LIF); breadth across incisive foramina canonical variates with the original cranial (BIF); length of diastema (LD); postpalatal measurements. All analytical procedures length (PPL); length of bony palate (LBP); were implemented using statistical routines breadth of bony palate (BBP); breadth of contained in Systat for Windows, Version (BZP); length of auditory 10.2 (2002). bulla (LAB); coronal length of maxillary tooth row (CLM); width of the upper first RESULTS AND COMPARISONS (WM1); depth of the upper (DI). INTRASAMPLE AGE AND SEX VARIATION: Five age classes were defined according to Superficial perusal of skulls in series of stage of eruption of the upper third molar western Mexican Oryzomys reveals substan- (M3) and relative wear of the upper molars tial range in size, shape, and pronouncement (M1–3): (1) juvenile—M3 not erupted, or if of cranial ridges, an impression remarked on erupted not reaching the occlusal plane; (2) by others who have worked with examples of immature—all molars fully erupted, little Oryzomys (Merriam, 1901; Goldman, 1918; wear on M1–2 but no wear on M3; (3) young Paradiso, 1960; Park et al., 1974; Humphrey adult—upper molars, including M3, lightly and Setzer, 1989). Some of this variation worn, enamel pattern retaining well defined plausibly corresponds to the age of individ- cusps and ridges, dentinal connections nar- uals, as suggested by occlusal wear of molars, row; (4) full adult—M1–3 moderately to but consistent size differences between the heavily worn with blunt cusps, wide dentinal sexes are also evident and seemingly vary connections, and enamel pattern variously independently of age. Such casual visual obscured; (5) old adult—occlusal wear so impressions are reinforced by statistical advanced that the enamel pattern is largely comparisons among age and sex cohorts obliterated, dentinal lakes of cusps are within our largest OTU of O. couesi mex- broadly continuous, and molars basinlike. icanus (N1, N 5 62). We use the abbreviations M1–3 or m1–3 to Many variables display regular, incremen- individually reference the upper (maxillary) tal increases in mean size across the three and lower (dentary) molars, respectively. adult age classes defined, producing age- Standard descriptive statistics (mean, correlated differences that contribute sub- range, standard deviation) were derived for stantially to nongeographic variation within the 11 OTUs using only those specimens our locality samples (table 1). The magnitude judged to be young, full, or old adult. of F values and their significance levels differ External data are provided as general guid- according to major anatomical regions of the ance to identification (table 1) but were not : most measurements taken on the facial subjected to morphometric comparisons. region (LR, BR, DR, LIF, BIF, LD), which Canonical variates (CVs) derived from mul- experiences substantial postweaning growth, tigroup discriminant-function classification disclose clear mean differences with greater and principal components (PCs) were com- age and achieve moderate significance levels, puted using only the 20 craniodental vari- whereas those dimensions measured on the ables, all of which were first transformed to neurocranium (BBC, BOC, DBC, IOB), natural logarithms. With only three speci- whose maturation plateaus earlier in postna- mens, the topotypes of lambi were entered as tal development, exhibit slight or no increases 98 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

TABLE 1 Arithmetic Means of External and Craniodental Variables and Results of One-way ANOVAs for Sex and Age Cohorts in Adult Oryzomys couesi mexicanus from Nayarit (OTU N1, n 5 62; Y, A, and O 5 young, full, or old adult age class, respectively; see Materials and Methods for variable abbreviations.)

Sex Age MF YA O Variable (N 5 31) (N 5 31) f(sex) (N 5 45) (N 5 13) (N 5 4) f(age)

TOTL 254.3 235.3 19.2*** 241.7 248.5 268.5 4.1* TL 129.9 120.3 15.1*** 123.2 128.5 136.0 3.7* HFL 31.0 30.0 15.5*** 30.5 30.4 31.0 0.4 ONL 30.1 28.8 12.5** 29.2 29.8 31.8 6.9** ZB 16.0 15.2 17.7*** 15.5 15.7 16.5 2.9 BBC 12.6 12.4 4.4* 12.5 12.4 12.6 0.4 BOC 6.8 6.7 3.8 6.7 6.7 6.8 0.5 DBC 9.5 9.3 7.7** 9.4 9.4 9.6 0.6 IOB 4.8 4.6 18.4*** 4.7 4.7 4.8 0.7 LR 9.5 8.9 11.8** 9.1 9.3 10.3 7.2** BR 5.8 5.5 11.6** 5.6 5.7 6.2 4.6* DR 6.6 6.3 9.6** 6.4 6.5 7.1 6.5** LIF 6.0 5.7 10.2** 5.8 6.0 6.2 3.7* BIF 2.3 2.3 1.9 2.2 2.3 2.5 7.3** LD 7.9 7.3 18.5*** 7.5 7.8 8.3 5.6** PPL 10.0 9.5 11.6** 9.6 9.8 10.7 5.0** LBP 5.5 5.1 16.7*** 5.3 5.2 5.8 4.5* BBP 5.5 5.4 3.8 5.5 5.5 5.6 1.5 BZP 3.3 3.1 5.9* 3.1 3.3 3.5 6.8** LAB 4.1 4.0 1.1 4.0 3.9 4.2 3.8* CLM 4.56 4.54 0.2 4.55 4.50 4.67 1.6 WM1 1.29 1.30 0.3 1.30 1.29 1.32 0.2 DI 1.85 1.79 3.8 1.77 1.89 1.99 8.7***

* 5 P # 0.05; ** 5 P # 0.01; *** 5 P # 0.001. with age and generally insignificant F values. and the generally smaller F values. The Dimensions of the molars (CLM, WM1), imbalance among age classes and predomi- which decrease in crown height with occlusal nance of young adults in OTU N1 typify age abrasion after eruption, do not grow in representation across all our analytical sam- length and width, unlike osseous components ples (young adults comprise ca. 70%, full of the skull; in contrast, the substan- adults 25%, and old adults only 5% of all tially increase in girth with age (table 1). In measured crania, respectively). nearly all of the age-sensitive variables, the Moderate to strong differences between significant differences reside between the the sexes were derived for most cranial young adult and full adult or old adult age measurements of Oryzomys couesi, males classes, but not between the full adult and old averaging 4%–8% larger than females de- adult age classes (according to Bonferroni pending upon the variable used as an index of post hoc multiple comparison tests). Such a size (table 1). Interestingly, molar dimensions pattern is predicated by sample sizes avail- (CLM, WM1) are among those few variables able for the three age groups and corre- that yielded no significant differences, a sponding mean differences. We suspect that finding that indicates that females possess the underrepresentation of the old adult more robust relative to the head cohort in OTU N1 accounts for the fewer skeleton. Humphrey and Setzer (1989), on variables that exhibit significant age effects the other hand, did obtain significant differ- 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 99 ences for molar measurements, but their biases in morphometric comparisons associ- variable definition was an alveolar length ated with imbalanced sex cohorts, we ran that would reflect changes in the bony palate three iterations of discriminant-function (ANOVAs of LBP were significant for age analysis (using only females, only males, and sex effects; table 1). Our next largest and both sexes) on three morphologically sample of O. couesi disclosed the same distinctive subspecies of O. couesi in western pattern of sexual and age variation, variable Mexico that vary in proportions of the sexes differences according to anatomical region, (albiventer, OTU J3 [4U,8-]; mexicanus, and attained significance levels (J1, N 5 60 or OTUs C2 [17U,3-], J2 [5U,15-], N1 [31U, 58; results not tabulated). The larger size of 31-]; peninsulae, OTU BC [10U,4-]). male Oryzomys was qualitatively noted by Comparable and unambiguous phenetic early researchers (Merriam, 1901; Goldman, structure among these five OTUs was dis- 1918; Paradiso, 1960) and quantitatively closed in bivariate plots of the first two CVs verified by Humphrey and Setzer (1989) in extracted in each iteration (fig. 2), and the subspecies of Oryzomys palustris. Further- percent variation explained by the first two more, growth studies of laboratory Oryzomys CVs is very similar in all three trials (81.1%– palustris have demonstrated that well-marked 82.4%). Moreover, the hierarchy of cranio- differences in weight and head-and-body dental phenetic resemblance among the five length of males and females become apparent OTUs, based on Mahalanobis’ distances by approximately 3–4 weeks in age (Park and between group centroids, is fully congruent. Nowosielski-Slepowron, 1972), or shortly The exact clustering distances and relative after weaning when dispersing subadults are disposition of OTUs on the first two CVs do entering the trappable population. The oc- differ according to the sexual subset ana- currence of strong sexual dimorphism in lyzed, but such minor departures would be Oryzomys contrasts with the pattern of expected as a result of the different numbers nongeographic variation reported for other of males and females available for each OTU oryzomyines, in which conspicuous age- and corresponding effects on covariance related size differences are demonstrable, matrices. but size disparity between the sexes is In summary, we combined the sexes and statistically trivial (e.g., —Carle- three adult age classes in the following ton and Musser, 1989; —Myers morphometric comparisons in order to am- and Carleton, 1981, Carleton and Musser, plify specimen numbers of the rarer taxa. 1995; —Voss, 1991). On aver- Although the contributions of secondary age, there is a slight female bias (56%) among sexual dimorphism and postweaning growth all adult skulls suitable for measurement, but (age) to within-sample variation are statisti- the proportion varies appreciably among cally appreciable, those sources of variation individual OTUs (33%–85% females). are typically negligible relative to the extract- Humphrey and Setzer (1989) procedurally ed factors that emerged as taxonomically circumvented the problematic variation in- informative among samples. Indeed, the troduced by consistent sexual size dimor- discriminant-function trials using one or the phism in O. palustris by confining their other or both sexes largely previewed the analyses to adult males, an approach achiev- results we obtained with the full data set. able with the large museum series available MORPHOMETRIC COMPARISONS: Pronounced for a common North American species. Such disparity in size and pelage color was especially selectivity proved impractical for several taxa evident among populations in Jalisco, where of interest here and would severely limit their key species-group taxa had been named sample sizes for morphometric comparisons. (albiventer Merriam, 1901; mexicanus Allen, Only four specimens of nelsoni and 21 of 1897), and initially sparked our suspicion peninsulae will ever be available for study; the about the possible existence of two species. 26 examples of albiventer that we managed to Two well-defined morphologies were visually assemble exemplify the relative rarity of this apparent in sorting locality samples: large- in North American collections of bodied, brightly colored specimens with Mexican Oryzomys. To explore possible broad skulls and robust molar rows versus 100 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Fig. 2. Three iterations of discriminant-function analysis of select OTUs using only females, only males, and the sexes combined. Left column, projection of specimen scores onto the first two canonical variates extracted; right column, phenograms based on Mahalanobis distances among group centroids. See text for discussion. 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 101 smaller individuals with narrower skulls and throughout the coastal lowlands of western more delicate molars (compare measure- Jalisco. ments for OTUs J1 and J3, table 2). Princi- We expanded our geographic perspective pal component ordination of all Jaliscan by including other samples of the broadly adults with intact skulls supplies additional ranging mexicanus and two forms with evidence of substantial craniodental differen- localized distributions in western Mexico, tiation over small geographic distances. nelsoni from the Tres Marı´as Islands and Projection of individual scores onto PCs I peninsulae from the Baja Peninsula. The only and II reveals two elliptical densities (fig. 3), analytically suitable locality series of albiven- one corresponding to specimens from the ter remains Merriam’s (1901) original type coastal plain and adjoining foothills in series (N 5 12) from Ameca, Jalisco. western Jalisco (5 mexicanus, ca. sea level Discriminant-function analysis of the 11 to 1000 m) and the other to those from OTUs representing these taxa disclosed three uplands in central and eastern Jalisco well-defined clusters in the first three canon- (5 albiventer, ca. 1200–1800 m). Although ical variates extracted (fig. 4), which collec- they marginally overlap, the major axes of tively summarize about 75% of the between- these elliptical spreads are phenetically dis- sample variation. The largest cluster consists crete (fig. 3), their Y-intercepts being signif- of the eight samples that represent mexicanus. icantly different between taxa (0.46 versus – Their centroids are closely approximate and 1.68; F 5 25.2, P , 0.001) while their slopes they broadly overlap one another in canon- are comparable (0.65 versus 0.41; F 5 1.4, P ical scores; a posteriori misclassifications 5 0.245). This covariation pattern among the among the eight is commonplace (24%–92% specimens of albiventer and mexicanus is a incorrect depending upon the OTU). The mexicanus aggregation is entirely separate familiar one that recalls multivariate con- from OTU J3 of albiventer along the first trasts recovered for other closely related, canonical variate, a hiatus mainly reflecting congeneric species of (Voss et al., the uniformly larger size of the latter 1990; Voss and Marcus, 1992; Carleton and (tables 2, 4), a distinction also emphasized Musser, 1995; Carleton et al., 1999). Predict- in the first component derived from PCA of ably, all variables load positively on the first all Jaliscan specimens. component and most correlations are mod- Although the sample of peninsulae is erate to large (r 5 0.55–0.98; table 3), intermediate to those of mexicanus and indicating a general size factor. Although albiventer in external size (table 2), cranial juveniles were omitted from our analyses, size and shape differences contribute to its some postweaning growth still accounts for sharp isolation from the latter OTUs along much of the elongate scatter observed within canonical variates 1–3 (fig. 4). Larger size in the species samples and for the oblique nearly all variables sets apart peninsulae from orientation of the group constellations on the various OTUs of mexicanus along the the first and second principal components. first canonical variate (correlations positive, Those variables that load heavily on PC II, coefficients moderately large; table 4). A more or less orthogonal to the major axes of subset of dimensions conveys certain shape the taxon samples, emphasize the robust differences that affect separation of peninsu- molars (CLM, WM1) of albiventer and, to lae along the next two latent variates, a lesser extent, the relatively broader expanse including a deeper braincase (DBC), broader of the incisive foramina (BIF) of mexicanus rostrum (BR), and expansive bony palate (tables 2, 3). The type specimen of albiventer (LBP, BBP). Noteworthy are the absolutely (USNM 82236) nests comfortably among and relatively broad incisive foramina (BIF) other large-bodied specimens from inland that characterize specimens of peninsulae, Jalisco (fig. 3). That of mexicanus (AMNH features that also contribute to the morpho- 2650/2128) was too damaged to include metric isolation of the Baja sample, especially in the PCA, but its small molars (CLM 5 on the second canonical variate (table 4). The 4.63 mm, WM1 5 1.26 mm) conform to incisive foramina in peninsulae even average the smaller Oryzomys found commonly wider than the large-bodied taxa albiventris 102 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

TABLE 2 External and Craniodental Measurements (in mm) of Samples of the Oryzomys couesi Complex in Western Mexico (Sample statistics include the mean, 6 1 SD, and the observed range; variable abbreviations are defined in Materials and Methods; column headings reflect our taxonomic recommendations.)

O. couesi mexicanus O. peninsulae O. albiventer O. nelsoni Jalisco Nayarit Baja California Jalisco Marı´a Madre Variable (J1, N 5 58) (N1, N 5 62) (BC, N 5 14) (J3, N 5 12) (TM, N 5 4)

TOTL 245.7 6 25.1 244.8 6 19.5 265.6 6 22.3 285.4 6 22.2 322 6 28.8 195–301 210–288 227–305 245–314 282–344 TL 132.3 6 13.2 125.1 6 10.7 136.8 6 10.8 155.4 6 14.3 181.5 6 14.6 102–160 105–150 114–156 129–173 160–191 HBL 113.3 6 12.8 119.7 6 11.0 128.9 6 12.3 130.0 6 8.2 140.5 6 14.8 91–141 96–144 113–152 116–142 122–153 HFL 30.2 6 1.7 30.5 6 1.2 32.0 6 1.2 36.1 6 1.9 37.3 6 1.7 26–34 27–33 29–34 33–40 35–39 ONL 29.4 6 1.4 29.5 6 1.5 31.5 6 1.7 32.9 6 1.3 34.8 6 2.5 26.4–32.5 26.7–33.2 27.8–34.3 30.0–34.5 31.7–37.8 ZB 15.8 6 0.9 15.6 6 0.8 17.1 6 0.9 17.6 6 0.6 18.0 6 1.0 13.8–18.0 13.9–17.9 15.1–18.7 16.6–18.3 16.6–18.9 BBC 12.5 6 0.4 12.5 6 0.4 12.9 6 0.2 13.2 6 0.3 13.4 6 0.3 11.7–13.5 11.7–13.3 12.4–13.3 12.8–13.8 13.0–13.6 BOC 6.7 6 0.3 6.7 6 0.2 6.8 6 0.2 7.1 6 0.2 7.2 6 0.3 6.1–7.3 6.3–7.2 6.2–7.0 6.9–7.6 6.8–7.6 DBC 9.1 6 0.3 9.4 6 0.4 9.9 6 0.4 9.7 6 0.3 10.0 6 0.2 8.3–9.8 8.4–10.5 9.4–10.8 9.3–10.1 9.9–10.2 IOB 4.7 6 0.2 4.7 6 0.2 4.9 6 0.2 4.8 6 0.2 5.2 6 0.3 4.1–5.1 4.3–5.3 4.7–5.3 4.5–5.1 4.8–5.6 LR 9.2 6 0.6 9.3 6 0.7 10.0 6 0.7 10.3 6 0.7 11.3 6 0.9 8.3–10.7 8.2–11.1 8.4–11.1 9.0–11.4 9.9–11.9 BR 5.8 6 0.4 5.6 6 0.4 6.4 6 0.5 6.1 6 0.3 7.0 6 0.6 5.0–6.7 4.8–6.9 5.4–7.4 5.4–6.5 6.2–7.5 DR 6.6 6 0.4 6.5 6 0.4 6.8 6 0.5 7.1 6 0.5 8.1 6 0.7 5.8–7.5 5.7–7.4 5 .7–7.7 6.2–7.7 7.0–8.7 LIF 5.8 6 0.3 5.9 6 0.4 6.4 6 0.5 6.5 6 0.4 6.4 6 0.3 5.0–6.5 5.0–6.7 5.5–7.3 5.7–6.9 6.1–6.9 BIF 2.2 6 0.2 2.3 6 0.1 2.6 6 0.2 2.4 6 0.2 2.5 6 0.1 1.9–2.7 2.0–2.7 2.3–3.2 1.9–2.6 2.4–2.6 LD 7.4 6 0.5 7.6 6 0.6 8.3 6 0.7 8.3 6 0.7 8.8 6 0.7 6.1–8.6 6.4–9.1 7.1–9.7 7.2–9.5 8.0–9.6 PPL 9.8 6 0.7 9.7 6 0.7 10.3 6 1.0 11.0 6 0.6 11.5 6 1.0 8.0–11.4 8.3–11.3 8.6–12.5 9.5–11.7 10.5–12.7 LBP 5.5 6 0.3 5.3 6 0.3 5.9 6 0.3 6.2 6 0.3 6.5 6 0.6 4.9–6.2 4.6–6.2 5.2–6.3 5.6–6.7 5.7–7.2 BBP 5.6 6 0.2 5.5 6 0.2 6.1 6 0.2 6.0 6 0.2 6.2 6 0.2 5.2–5.9 5.1–6.0 5.5–6.6 5.7–6.2 5.9–6.4 BZP 3.1 6 0.2 3.2 6 0.3 3.3 6 0.3 3.7 6 0.2 4.2 6 0.3 2.6–3.6 2.6–3.8 2.7–3.9 3.5–4.0 3.7–4.5 LAB 4.0 6 0.1 4.0 6 0.2 4.1 6 0.2 4.5 6 0.2 4.4 6 0.3 3.7–4.4 3.5–4.4 3.9–4.5 4.2–4.8 4.1–4.6 CLM 4.60 6 0.14 4.56 6 0.17 4.72 6 0.11 5.24 6 0.19 4.92 6 0.12 4.23–4.94 4.22–4.91 4.53–4.94 4.74–5.51 4.75–5.00 WM1 1.35 6 0.05 1.30 6 0.07 1.36 6 0.05 1.53 6 0.06 1.52 6 0.04 1.21–1.45 1.04–1.40 1.26–1.44 1.45–1.64 1.47–1.56 DI 1.75 6 0.12 1.81 6 0.14 2.18 6 0.21 2.06 6 0.13 2.31 6 0.18 1.53–1.98 1.54–2.19 1.83–2.59 1.83–2.26 2.06–2.47 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 103

Fig. 3. Projection of specimen scores onto first two axes extracted from principal component (PC) ordination of 20 log-transformed variables as measured on intact skulls (N 5 108) representing Oryzomys couesi albiventer and O. c. mexicanus from Jalisco, Mexico. The position of the type specimen (holotype) of albiventer Merriam (USNM 82236) is indicated by ‘‘ht.’’ See table 3 for variable correlations and variance explained. and nelsoni (table 2). Clustering (UPGMA) Variable loadings that influence the isola- based on Mahalanobis’ distances between tion of albiventris similarly apply to the OTU centroids does indicate a closer phenetic of nelsoni, insofar as the first three variates affinity between peninsulae and OTUs of extracted are concerned (fig. 4). Compared mexicanus, compared with albiventris and with examples of mexicanus and peninsulae, nelsoni, and simultaneously its strong differ- specimens of nelsoni and albiventris are large entiation from any one of them (fig. 5). In Oryzomys that possess more robust skulls principal-component and discriminant-func- and notably heavier molars (table 2), differ- tion analyses confined to samples of peninsu- ences evidenced in the loadings on canonical lae and mexicanus (not illustrated), contrast variates one and two (table 4). Clear-cut in incisor girth (DI) emerges as an important circumscription of the four specimens of variable for separation of the two taxa nelsoni from those of albiventer, however, is (table 4), in addition to those already men- captured on the fourth and fifth variates (not tioned. No specimen of peninsulae was plotted), which principally emphasize the mistakenly associated with another OTU massive development of the rostrum (LR, according to post hoc probabilities of group WR, DR, BZP) and upper incisors (DI) in membership. nelsoni (tables 2, 4). Given the scope of taxa 104 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

TABLE 3 type specimen of aztecus (USNM 51173) Correlations (Loadings) of 20 Log-transformed could not be measured. Type material of the Craniodental Variables with Derived Principal three species-group taxa that originate within Components (PC) Extracted from Ordination coastal plain were all definitely assigned to Using Intact Specimens (N = 108) of Oryzomys couesi albiventer and O. c. mexicanus mexicanus (fig. 6), including the holotypes of in Jalisco, Mexico bulleri Allen (AMNH 6663 to N3) and rufus (See fig. 3; variable abbreviations are defined Merriam (USNM 91404 to N2), as well as the in Materials and Methods.) three original topotypes of O. couesi lambi Burt (UCLA 51201 to N2, UCLA 51026 to Variable PC I PC II C2, UCLA 51038 to J1); none demonstrated ONL 0.98*** 0.03 even a low level of affinity for the OTUs of ZB 0.94*** 20.02 albiventer or peninsulae. Those taxa described BBC 0.75*** 20.09 from upland localities in central Mexico were BOC 0.71*** 20.05 variously affiliated (fig. 6). The specimen of DBC 0.84*** 20.03 molestus Elliot (FMNH 8667), described from IOB 0.59*** 0.19 Ocotla´n, Jalisco, was dispersed just peripheral LR 0.93*** 0.15 to the type series of albiventer (Ameca, BR 0.87*** 0.28 DR 0.92*** 0.18 Jalisco) but definitely classified with that LIF 0.87*** 0.12 OTU (P 5 1.000); the type of albiventer BIF 0.65*** 0.40** Merriam (USNM 82236) naturally clustered LD 0.94*** 0.15 among the original type series. Type speci- PPL 0.93*** 0.18 mens of aztecus Merriam (USNM 51173, LBP 0.81*** 20.25 Yautepec, Morelos) and regillus Goldman BBP 0.84*** 20.22 (USNM 125945, Los Reyes, Michoaca´n) BZP 0.91*** 20.18 plotted at the margin of our large represen- LAB 0.68*** 20.26 tation of mexicanus, but their canonical scores CLM 0.64*** 20.59*** statistically aligned them with those OTUs WM1 0.55*** 20.65*** DI 0.87*** 20.11 (the greatest affinity of both with J2), not with Eigenvalues 0.069 0.006 albiventer. Finally, the holotype of crinitus % variance 72.2 6.3 Merriam (USNM 50182, Tlalpan, Distrito Federal) was associated with the series of *** P # 0.001; ** P # 0.01. peninsulae, at an unexpectedly strong level of probability (P 5 0.996). The large values under consideration, covariation among measured for the incisive foramen width these variables emerged as minor compared (BIF), rostral breadth (BR), and braincase with the overarching patterns of between- depth (DBC) of this specimen (table 5), group variation summarized by the first three variables that importantly contribute to the canonical variates. Nonetheless, their unique discrimination of mexicanus and peninsulae and consistent expression in nelsoni accounts (see above), presumably influenced its multi- for its strong differentiation in the pheno- variate disposition and a posteriori classifica- gram generated from Mahalanobis’ distances tion in peninsulae. between OTU centroids (fig. 5). The weak clustering association obtained between nel- DISCUSSION soni and albiventer approximates that derived for peninsulae and OTUs of mexicanus. TAXONOMIC RECOMMENDATIONS AND Thosetypespecimenswithintactskullswere FUTURE STUDY: We recognize four species entered as unknowns in a 10-group discrimi- among our study samples from western Mex- nant-function analysis for classification ac- ico: Oryzomys albiventer Merriam, O. couesi cording to a posteriori probabilities of group mexicanus J.A. Allen, O. nelsoni Merriam, and membership. The OTU of O. nelsoni was O. peninsulae Thomas (figs. 7–9). These forms excluded as a classificatory option; further- can be objectively defined by a combination of more, DI was omitted in calculating discrim- qualitative morphological traits, demonstrate inant coefficients because the incisors in the cohesive patterns of morphometric identity, 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 105

Fig. 4. Results of 11-group discriminant function analysis based on 20 log-transformed measurements of intact specimens (N 5 246) representing the Oryzomys couesi group from western Mexico: top, projection of specimen scores on canonical variates (CV) 1 and 2; bottom, projection of specimen scores on canonical variates 2 and 3. Individual scores are plotted for the four known specimens of O. nelsoni; minimum convex polygons enclose individual scores around the centroid for OTUs of albiventer and peninsulae. To avoid visual clutter only centroids are indicated and a single bounding polygon encloses maximum dispersion for all specimens of mexicanus. See table 4 for variable correlations and variance explained. 106 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

TABLE 4 Correlations (Loadings) of 20 Log-tranformed Craniodental Variables with Derived Canonical Variates (CV) (Based on 11-group discriminant function analysis of intact specimens, N 5 246, representing Oryzomys in western Mexico; see fig. 4 for scatter plots of CVs 1–3. Variable abbreviations are defined in Materials and Methods.)

Variable CV 1 CV 2 CV 3 CV 4 CV 5

ONL 0.69*** 20.07 20.09 20.22 0.27** ZB 0.65*** 20.09 20.22 20.19 0.15 BBC 0.61*** 20.07 20.05 0.09 0.25** BOC 0.24* 20.11 20.01 0.13 0.20 DBC 0.65*** 0.29*** 0.04 0.05 0.13 IOB 0.46*** 0.17 20.03 20.33*** 0.27** LR 0.59*** 20.11 20.14 20.16 0.28** BR 0.46*** 20.05 20.38*** 20.42*** 0.29*** DR 0.55*** 20.16 20.06 20.33*** 0.43*** LIF 0.62*** 0.05 20.15 20.25** 0.01 BIF 0.56* 0.35*** 20.22* 20.23* 0.16 LD 0.62*** 0.05 20.12 20.23* 0.12 PPL 0.51*** 20.11 20.04 20.35*** 0.15 LBP 0.65*** 20.29*** 20.31*** 0.00 0.25** BBP 0.70*** 20.05 20.37*** 20.28** 0.16 BZP 0.62** 20.08 0.12 20.44*** 0.13 LAB 0.62*** 20.21 0.03 20.21 0.03 CLM 0.76*** 20.40*** 20.09 20.11 20.12 WM1 0.62*** 20.56*** 20.23* 20.07 0.00 DI 0.75*** 0.20 20.15 20.16 0.24** Eigenvalues 2.19 1.66 0.91 0.73 0.41 Canonical Correlations 0.83 0.79 0.69 0.65 0.54 % variance 34.5 26.1 14.2 12.4 6.4

*** P # 0.001; ** P # 0.01; * P # 0.05. and occupy discrete geographic ranges (see mal study since their descriptions; they are not emended diagnoses, distributions, and re- ‘‘cryptic’’ species. marks under Taxonomic Summary). Such As now understood (i.e., Weksler et al., well-marked contrasts stand at odds with the 2006, plus species recognized herein), Oryz- conventional notion of intergrading subspe- omys is a moderately speciose genus (8 cies as conveyed in past studies and influential species to date considered valid). The distri- classifications (Goldman, 1918; Hall, 1981). bution of most species is confined to Neo- The genesis of the widely distributed, polytyp- tropical environments of Middle America, ic view of O. palustris, encompassing couesi but one inhabits temperate settings in south- and most other Middle American taxa, hinged eastern North America (O. palustris) and on a geographically and taxonomically limited another occupies the northernmost margin of study (Hall, 1960) that was uncritically South America (O. gorgasi). Even so recon- expanded in the latter 1900s (Handley, 1966; stituted, Oryzomys still contains several Hershkovitz, 1971; Hall, 1981), compelled species-level uncertainties. The widely rang- more by the revisionary habit of the era than ing, polytypic definition of Oryzomys palus- by fresh analyses of character data. Consid- tris, as classically formulated in Hall (1981), ered in this historical context, the sharp reminds us of other Middle American species differentiation we encountered among western whose composite nature has been emphati- Mexican taxa, unexamined in any systematic cally documented by recent researches. For detail since Goldman’s (1918) revision, is example, Oryzomys rostratus has been sepa- unsurprising. These taxa have received mini- rated from O.‘‘melanotis’’ (Engstrom, 1984); 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 107

Fig. 5. Pattern of phenetic similarity among the 11 OTUs of Oryzomys from western Mexico based on UPGMA clustering of the squared Mahalanobis distances among group centroids as derived from discriminant-function analysis (see fig. 4). Currently recognized species-group taxa are shown above the appropriate stem subtending terminal OTUs.

Sigmodon arizonae, S. hirsutus, S. mascoten- land coast, where a related species [i.e., O. sis, and S. toltecus have been recovered from couesi mexicanus] is plentiful.’’ His conjecture S.‘‘hispidus’’ (Zimmerman, 1970; Carleton et on the nonnative status of the peninsular al., 1999; Peppers and Bradley, 2000; Peppers form may have influenced Hall (1981) to et al., 2002; Carroll et al., 2005); and arrange it within a polytypic O. palustris. Peromyscus aztecus, P. beatae, P. levipes, P. However, the morphological distinctiveness madrensis, P. sagax, P. simulus,andP. and morphometric separation of peninsulae spicilegus have been resurrected as valid that we observed are substantial and instead species, all of them formerly masked under portray a species that evolved in situ, isolated P.‘‘boylii’’ (Carleton, 1977, 1979; Carleton et appreciably longer from the mainland distri- al., 1982; Rennert and Kilpatrick, 1987; bution of Oryzomys than a recent anthropo- Schmidly et al., 1988; Bradley et al., 1989, genic introduction would suggest. 1996; Castro-Campillo et al., 1999). In The phenetic linkage of O. couesi mexica- similar fashion, future taxonomic research nus and O. peninsulae offers a preliminary upon Oryzomys, in particular the O. couesi kinship hypothesis to be tested with other section, will undoubtedly disclose greater data, and two biogeographic scenarios de- systematic diversity, and some of those serve consideration in reconstructing the research avenues might explore or solidify latter’s relict distribution. Oryzomys peninsu- the following points: lae may have differentiated from an ancestral (1) The remoteness of the Oryzomys stock whose distribution once covered the population that once existed at the tip of Baja Peninsula, extreme southwestern United Baja California, far removed from the core States, and western Mexico. Several rodent generic distribution in mainland Mexico, species or species complexes today exhibit a prompted Nelson (1922: 124) to remark that similar distributional pattern (e.g., see ranges it ‘‘may readily have been introduced in in Lawlor, 1983: appendix 10), suggesting shipments of bananas or other farm products that Pleistocene climatic changes and subse- from Mazatlan or elsewhere along the main- quent range contraction could have isolated 108 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Fig. 6. Dispersion of type specimens of Oryzomys named from western Mexico based on a posteriori classification using discriminant coefficients extracted from 10-group discriminant function analysis of albiventer, couesi mexicanus,andpeninsulae. Minimum convex polygons enclose individual scores around the group centroid (or grand centroid in the case of couesi mexicanus). See text for discussion. an Oryzomys population at the tip of the ago) and accelerated seafloor spreading peninsula. The extension of O. couesi mex- (beginning approximately 3.5 million years icanus along the coastal plain in western ago) eventually brought the insular cape Mexico, as far north as west-central Sonora block to its present peninsular terminus, a (fig. 7), further intimates the former continu- geological nonconformity demarcated by ity of Oryzomys populations on the mainland the Isthmus of La Paz. The biotic legacy of surrounding the Sea of Cortez. the former juxtaposition of the southern Alternatively, O. peninsulae may have cape block and west-central Mexico is originated when the southernmost cape suggested by the current distributions of region of the Baja Peninsula was positioned certain plants and birds (Cody and Velarde, as an island approximate to the Nayarit- 2002; Cody et al., 2002). Phylogeographic Jalisco coast in the late Miocene (Carren˜o analysis, using ancient DNA, may disclose and Helenes, 2002). Geologic and tectonic sufficiently fine-scale genetic affinities to evidence indicates that northwestward rifting choose among these or other biographic (beginning approximately 5.5 million years scenarios. 09CREO N ARROYO-CABRALES: AND CARLETON 2009 TABLE 5 Measurements a (mm), Sex, and Age of Type Specimens of the Oryzomys couesi Group in Western Mexico (Y, A, and O 5 young, full, or old adult age class, respectively. Variable abbreviation s are defined in Materials and Methods.)

O. albiventer O . nelsoni O . peninsulae O . couesi mexicanus O .c.f. couesi albiventer molestus b nelsoni peninsulae mexicanus bulleri rufus lambi c crinitus aztecus regillus USNM FMNH USNM BMNH AMNH AMNH USNM UCLA USNM USNM USNM 82236 8667 89200 98.3.1.107 2650/2128 8359/6663 91404 topotypes 50182 51 173 125945

Sex M M M M M M F — M M M AgeA O A A A Y O —YYA TOTL 288 325 342 298 249 242 250 226–229 307 290 305 TL 153 170 190 190 142 127 136 111–116 161 154 169 HFL 37.5 38 38 34 30 27 28 29–30 37 35 36 EL — 15 — 18 13 11 — 17 — — — ONL 32.7 36.1 37.7 33.8 — 28.0 28.1 30.2–30.4 32.6 32.3 33.5 ZB 17.9 19.4 18.6 18.3 — 14.9 15.9 15.9–16.5 17.0 16.9 18.2 BBC 13.7 14.1 13.3 13.2 12.6 12.7 12.2 12.1–12.4 13.4 13.0 13.6 BOC 7.2 7.7 7.2 — — 6.4 6.4 6.7–6.8 6.9 7.2 7.2 DBC 10.2 10.6 9.9 — — 8.7 9.3 8.8–8.9 10.6 9.9 10.1 IOB 5.0 5.5 5.5 5.2 4.7 4.5 4.8 4.6–4.7 5.0 5.0 5.0 LR 10.6 11.8 12.4 9.7 — 8.9 8.6 9.6–9.8 10.3 10.4 10.9 RZMSCOUESI ORYZOMYS BR 6.4 7.2 7.2 7.1 6.3 5.1 5.7 5.8–6.1 6.5 6.3 6.9 DR 7.3 8.1 8.7 — — 5.9 6.5 6.5–6.6 7.3 7.2 7.5 LIF 6.9 7.8 6.9 7.0 6.4 5.3 6.3 6.0–6.3 7.0 6.7 6.7 BIF 2.6 2.2 2.4 2.9 1.5 1.8 2.3 2.1–2.4 2.8 2.6 2.5 LD 8.8 10.1 9.6 8.8 8.2 6.8 7.5 8.1–8.5 9.0 8.5 9.0 PPL 11.1 12.9 12.7 12.1 — 8.9 9.3 10.1–10.3 11.3 11.2 11.6 LBP 5.9 6.6 7.2 6.0 5.7 4.8 4.9 5.1–5.4 5.6 6.0 6.3 BBP 6.2 — 6.0 6.5 5.6 5.4 5.3 5.6–5.8 5.9 5.9 6.2 BZP 3.5 3.9 4.3 3.6 3.5 2.7 3.2 3.2–3.3 3.1 3.2 3.7

LAB 4.5 4.9 4.4 4.4 4.2 3.9 4.3 4.2–4.6 4.8 4.3 4.5 109 COMPLEX CLM 5.17 5.60 4.96 5.22 4.63 4.46 4.24 4.32–4.77 4.87 4.97 4.77 WM1 1.58 1.63 1.56 1.53 1.26 1.28 1.26 1.40–1.46 1.44 1.44 1.44 DI 2.01 2.02 2.43 — 2.07 1.76 1.74 1.76–1.89 1.84 — 1.95

a External measurements as obtained from skin tags or given in original des criptions; craniodental measurements taken by the authors. b Craniodental measurements supplied by W.T. Stanley, FMNH. c Based on three original topotypes identified in Burt’s (1934) subspecie s description (UCLA 51021, M, YA; UCLA 51026, F, YA; UCLA 51038, F, A). 110 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Fig. 7. Distribution of four species of Oryzomys here recognized in western Mexico, based on specimens examined by us. Gray shading represents the same elevation zones depicted in figure 1.

(2) The Oryzomys population that once moved from the coastal lowlands where the inhabited Isla Marı´a Madre, the largest of progenitor of the island species must have the Tres Marı´as archipelago off Nayarit, has once occurred. The two were paired as the been accepted as a species by most authors phenetically most similar OTUs in all multi- since Merriam (1898) first described it as O. variate analyses of craniodental variables nelsoni. In spite of the small sample size (only (figs. 4, 5), and this linkage provides a four specimens known), the differentiation of working hypothesis of relationship to be O. nelsoni is sharply circumscribed from tested with other data. Peromyscus madrensis, mainland forms and sustains its continued another rodent endemic of the Tres Marı´as recognition as a species, following Merriam Islands, was apparently isolated early during (1901) and Goldman (1918) and contra the radiation of the P. boylii species group Hershkovitz (1971). In retaining nelsoni as a (Tiemann-Boege et al., 2000). In like manner, species, Goldman (1918) considered it to be perhaps O. nelsoni represents an old lineage closely allied to the nearest mainland form, within the diversification of western Mexican O. couesi mexicanus. However, by its size, O. Oryzomys, diverging from the common nelsoni invites careful comparison with O. ancestor that also gave rise to O. albiventer. albiventer, a species whose distribution is (3) We elected a conservative course and currently confined to the interior plateau retained mexicanus Allen as a subspecies of region of west-central Mexico (fig. 7), re- O. couesi in lieu of the detailed analyses and 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 111 34.6 5 tla, near (USNM 89201, ONL nelsoni . O , B 32.3 mm), a young adult male from Baja 5 0 ft; representing species here recognized from western Oryzomys 32.2 mm), a young adult male from Nayarit, 1 mi S Cuau 5 (USNM 146610, ONL )ofadult 3 peninsulae . O , C views (ca. 2 (USNM 509418, ONL 33.8 mm), a young adult male from Jalisco, Ameca, 400 a Madre, 800 ft; 5 ´ couesi mexicanus . O , D (USNM 82240, ONL albiventer . O , A Fig. 8. Dorsal (upper row) and ventral (lower row) cranial sea level. mm), an adult femaleCalifornia Sur, from San Nayarit, Jose´del Cabo; Isla Marı Mexico: 112 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Fig. 9. Lateral cranial views (ca. 2 3) of the same specimens adult Oryzomys illustrated in figure 8: A, O. albiventer; B, O. nelsoni; C, O. peninsulae; D, O. couesi mexicanus. broader geographic canvas required to clarify isthmus (Carleton et al., 1999; Peppers and its status. Visual comparisons of skins and Bradley, 2000). No such morphological skulls suggested that, among the four species dichotomy is apparent to us based on in western Mexico, examples of mexicanus inspection of O. couesi specimens from the most closely resemble those of O. couesi eastern lowlands of Mexico, the Isthmus of proper, whose type locality is Coban, Gua- Tehuantepec, and Guatemala, including temala. Recent studies of Sigmodon ‘‘hispi- those that belong to the currently recognized dus’’ and Peromyscus ‘‘boylii’’ cast suspicion subspecies couesi and zygomaticus (e.g., Hall, on such a conclusion: subspecies distributed 1981). Other data sources and a broader along the western versus eastern versants of geographic perspective are needed, one that Mexico in those so-called species have been stresses additional sampling in the Isthmus of demonstrated to be distinct from one another Tehuantepec where couesi mexicanus and O. (see above references). In particular, the c. couesi are thought to intergrade (Goldman, distribution of O. couesi mexicanus is exten- 1918; Hall, 1981). Acquisition of topotypic sively congruent with that of Sigmodon material from the type localities of the several mascotensis (Carleton et al., 1999), its range forms in eastern and southern Mexico placed penetrating the interior along the basin of the in full synonymy under O. c. couesi—jalapae Rı´o Balsas and following the coastal plain to Allen and Chapman (1897), jalapae rufinus the Isthmus of Tehuantepec. In this case, Merriam (1901), jalapae apatelius Elliot however, examples of the large-bodied S. (1904), goldmani Merriam (1901), teapensis mascotensis are readily distinguishable from Merriam (1901)—may prove critical. the smaller S. toltecus where their ranges Farther south, the status and relationships approach one another in the region of the of several other epithets will bear on delim- 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 113 iting the southern distributional occurrence craniodental variables, associates them with of O. couesi proper. Jones and Engstrom O. couesi mexicanus (aztecus and regillus)or (1986) considered richmondi Merriam (1901), with O. peninsulae (crinitus), not with O. whose type locality is in southeastern Nicar- albiventer. The holotypes and type series of agua (Zelaya, 50 mi above Bluefields), to be aztecus and regillus consist of large individ- another that is inseparable from O. uals whose dimensions measure at the upper couesi couesi. Our examination of that type limits of variation observed within typical O. (USNM 36340/48705) indicates that it is an couesi mexicanus (compare tables 2 and 5), example of O. couesi, but evaluation based and it is zoogeographically plausible that on denser geographic sampling is necessary these names simply apply to more robust before one might comment on subspecific populations of the same species where they divisions. Another Oryzomys described from occupy higher elevations. Whether they southeastern Nicaragua (Zelaya, 7 mi below represent yet another species within the O. Rama) is the enigmatic O. dimidiatus Thomas couesi group must await improved sample (1905), known by only two specimens, but sizes and additional data sources to discern investigators who have examined the taxon broader patterns of variation and critically have retained it as a valid species (Genoways decide their status. We do not accept that and Jones, 1971; Jones and Engstrom, 1986; populations represented by the type of Musser and Carleton, 1993; Sa´nchez et crinitus, described from over 2000 m in the al., 2001). In reporting the rediscovery of Valley of Mexico, and the series of peninsu- O. gorgasi Hershkovitz (1971), Sa´nchez lae, described from near sea level at the tip of et al. (2001) confirmed its distinction from Baja California, form one and the same other forms described from southern Cen- biological species. As in the case of aztecus tral America, including dimidiatus Thomas and regillus, we view these discriminant- (1905), gatunensis Goldman (1912), and function results as indicative of taxonomic richmondi Merriam (1901), the latter two problems in need of resolution and the taxa now placed under O. couesi. The probable existence of greater species diversity distribution of O. gorgasi plausibly extends than offered by our predefined OTUs. The into southern Central America, and museum recent photograph of a large, long-tailed series from this region should be reexamined Oryzomys from the Distrito Federal (Medel- with the fresh perspective supplied by Sa´n- lı´n and Medellı´n, 2006: 710), verifies that chez et al. (2001). populations referable to crinitus Merriam still (4) Within his arrangement of O. couesi, persist in the Valley of Mexico and invite Goldman (1918: 37) identified a complex of elucidation of their systematic relationships. related subspecies from central Mexico: ‘‘It (5) Oryzomys fulgens Thomas (1893) is a [O. c. crinitus] is closely allied to the other little-known and overlooked taxon, but forms inhabiting river valleys of the plateau clarification of its status will prove pivotal region of Mexico—O. c. aztecus, O. c. in deciding nomenclatural issues that concern albiventer,andO. c. regillus.’’ As argued albiventer and the other upland taxa, espe- above, we acknowledge O. albiventer to be cially crinitus. Thomas described fulgens specifically distinct from O. couesi mexicanus, based on one of the three specimens origi- the form distributed along coastal plain and nally used by Alston (1876) in his character- lower mountain slopes in western Mexico ization of Hesperomys couesi. In doing so, (fig. 7). Of the four taxa described from Thomas selected a lectotype for H. couesi interior uplands, O. albiventer Merriam (specimen ‘‘a’’ 5 BMNH 75.2.26.15) and (1901: 279) is an older name than regillus fixed its type locality (Coban, Guatemala). Goldman (1915) and has page priority over The type locality of the specimen that he crinitus Merriam (1901: 281) and aztecus designated as the type of O. fulgens (‘‘b’’ 5 Merriam (1901: 282). Evidence for the BMNH 70.6.20.3), collected by the French synonymy of these forms with O. albiventer naturalist Adolphe Boucard (1839–1905), is so far unpersuasive, however. was given only as ‘‘Mexico, exact locality Post hoc classification of the type speci- unknown.’’ Thomas supposed that it origi- mens of the latter three taxa, based on 19 nated from ‘‘southern Mexico, probably in or 114 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331 near the Valley of Mexico,’’ and his suppo- Hall, 1981). As demonstrated for O. antillarum sition about the source of Alston’s specimen by Morgan (1993) and documented herein for ‘‘b’’ has been repeated in the subsequent O. nelsoni and O. peninsulae, however, the taxonomic literature on Oryzomys (Merriam, differentiation of these insular and peninsular 1901; Goldman, 1918; Miller and Kellogg, forms is appreciable and contradicts their 1955; Hall, 1981). No one has since offered a continued maintenance under the one main- formal restriction of the type locality, how- land species O. couesi. Other insular and ever. In spite of the possible origin of fulgens peninsular forms (azuerensis, cozumelae, and from the Valley of Mexico, Merriam (1901) gatunensis) now retained within O. couesi proceeded to name another species, O. similarly deserve objective assessment of their crinitus, from the same region (Distrito status. The generic assignment of the subfossil Federal, Tlalpan) and suspected that it ‘‘will form described from Curac¸ao (McFarlane and prove only a subspecies of fulgens.’’ Merriam Debrot, 2001), a continental-shelf island in the had not examined the type of fulgens but Netherlands Antilles, and its possible synon- based his impression of their close similarity ymy with O. gorgasi also warrant attention on Thomas’ description, which he quoted in (see comments in Musser and Carleton, 2005: full. Apparently the indeterminate type lo- 1148, and Weksler et al., 2006: 4). cality of fulgens persuaded Merriam to name CONSERVATION NOTES: Oryzomys nelsoni crinitus as a full species based on a well- is known by only four specimens of the prepared specimen newly collected by Nelson original type series (Merriam, 1898), ob- and Goldman, conveniently at hand, and tained by Nelson and Goldman from Marı´a with undoubted provenience. Madre Island in May 1897, and is considered When nearly all Middle American forms extinct (Ceballos et al., 2002; Alvarez-Casta- of Oryzomys were viewed as races of the n˜eda and Me´ndez, 2003; IUCN, 2006). The older taxa O. couesi (Alston, 1876) or O. specimens were collected in moist herbaceous palustris (Harlan, 1837), the imprecise geo- vegetation surrounding springs located on graphic source of the junior name fulgens the upper slopes (, 245 m) of Marı´a Madre Thomas (1893) was nomenclaturally incon- (Nelson, 1899a, 1899b), the largest of the sequential. Its possible senior status among Tres Marı´as Islands. No other dedicated Oryzomys taxa named from uplands in inventory of small occurred until interior Mexico now invites careful archival Don E. Wilson led a Biological Survey field research on Boucard’s travels to see whether team to the Tres Marı´as in March 1976. a definite type locality can be established and Wilson and colleagues located the area the name can be critically applied to Oryz- described by Nelson where O. nelsoni had omys populations. Having satisfied that been discovered but collected only crucial step, the relationship of fulgens to rattus; Wilson (1991) inferred that the abun- albiventer, crinitus, and aztecus can be dant presence of the commensal may have addressed and synonymies allocated if ap- contributed to the extinction of O. nelsoni. propriate. Another Tres Marı´as endemic rodent, Per- (6) Individuals of Oryzomys commonly omyscus madrensis, was recovered in moder- inhabit coastal wetlands (Sharp, 1967; Wolfe, ate numbers on Isla Marı´a Madre along with 1982, 1985), are adept swimmers (Svihla, 1931; Rattus rattus (Carleton et al., 1982; Wilson, Hamilton, 1946; Esher et al., 1978), and are 1991). known to emigrate between near-shore islands Current evidence holds little promise that (Forys and Dueser, 1993). Their semiaquatic populations of O. peninsulae persist. The habits and ecological preference readily ex- form is documented from only two localities, plain the common presence of Oryzomys on Santa Anita and San Jose´ del Cabo, in continental-shelf islands and perhaps the one southernmost Baja California. The original oceanic occurrence (O. antillarum on Ja- series (6 specimens) from Santa Anita, the maica). Several of these peripheral isolates type locality (Thomas, 1897), was collected acquired species-group epithets, all of which by Dane Coolidge (1873–1940), a freelance were hastily reallocated as subspecies of the field collector and writer, in May 1896. The all-inclusive O. palustris (Hershkovitz, 1971; 15 USNM specimens from San Jose´ del Cabo 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 115 were obtained by Nelson and Goldman in Jalisco. On the other hand, we encountered January 1906; the companions afterward few examples of O. albiventer in North traveled to Santa Anita but failed to collect American collections, only 26 specimens any topotypic examples of O. peninsulae.The representing 11 localities. Most of these localities are about 13 km apart and both represent material collected by Nelson and were situated along the Rı´o San Jose, a Goldman in Jalisco over the years 1897– permanent river around the turn of the 1903, supplemented by scattered specimens century that flowed through the predomi- obtained by KU field crews over the years nantly arid landscapes in southern Baja 1949–1969. The distribution of O. albiventer California (Nelson, 1922; Goldman, 1951). largely conforms to the sloping plateau, ca. Goldman (1918: 46) reported that ‘‘The 1200–1900 m, surrounding Lago de Chapala range of the species seems to be centered in and its greater drainage basin. Given its the marshes near the mouth of the San Jose´ hydrographic advantages and more mesic River at San Jose´ del Cabo, extending a few climate, the region has long attracted large- miles inland along the course of the stream,’’ scale agricultural development, and Goldman and that ‘‘The neighboring coast is extremely (1951) described the dramatic environmental arid and unsuitable for habitation by an changes that had transpired between his early Oryzomys.’’ Subsequent habitat changes over (1902–1903) and later visits (1926, 1935) to the latter half of the 20th century in southern the area (see especially accounts for La Barca Baja California have been dramatic and and Ocotla´n). Although populations must be would be inimical for a semiaquatic rodent highly fragmented, we believe that the species tied to such a precarious ecological niche. survives in pockets of appropriate habitat. The river no longer exists, having been Efforts should be devoted to surveying the sapped dry for irrigation projects, and the current status of O. albiventer, including marshy lagoon at its mouth in San Jose´ del refinement of its distribution based on Cabo is polluted commensurate with devel- vouchered specimens contained in Mexican opment of the shoreline for tourism (Alvarez- collections. This refinement necessarily in- Castan˜eda, 1994). Two deliberate attempts to volves resolution of its relationship to the rediscover peninsulae in the late 1900s proved other forms described from central Mexico unsuccessful: Michael A. Bogan and a Bio- (see topics 4 and 5 above). logical Survey field team visited San Jose´ del Cabo for one intensive night of collecting in TAXONOMIC SUMMARY 1979, and Ticul Alvarez-Castan˜eda (1994) surveyed habitats in and around both Santa Synonymies are presented below for the Anita and San Jose´ del Cabo several times four species of Oryzomys recognized in over 1991–1993. The vulnerability of O. western Mexico and trace their earliest peninsulae has been obscured by its confusion identification and first subsequent usage in under O. couesi, a species viewed as broadly other name combinations. Emended diagno- distributed and accordingly classified as ses are presented for those two (O. albiventer Lower Risk—least concern (IUCN, 2006). and O. peninsulae) that have long been The of O. peninsulae must confused with O. palustris or O. couesi (e.g., be considered critically endangered, if not fide Goldman, 1918, and Hall, 1981). We extinct, in view of its lack of documentation in regard the status of aztecus Merriam, crinitus over a century, its restricted geographic Merriam, and regillus Goldman from central occurrence, and the environmental deteriora- Mexico as unresolved and provisionally tion that has befallen its preferred habitat. retain them within O. couesi sensu lato Of the two remaining species we identify (appendix 1). Specimens examined include within western Mexico, populations of O. all individuals personally seen and identified couesi mexicanus appear to deserve the IUCN by us. An emended diagnosis is also given for designation of ‘‘Lower Risk—least concern.’’ the phylogenetically restricted construct of Large series of this species have been Oryzomys. See Weksler et al. (2006) for documented within the latter 1900s, e.g., character traits and diagnoses of the new from lowlands of Nayarit, Colima, and genera that have been removed from the 116 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

polyphyletic shell that so long characterized DISTRIBUTION: Palustrine, lacustrine, and ‘‘Oryzomys’’ in its traditional usage (Mer- riverine habitats in the southeastern United riam, 1901; Goldman, 1918; Hall, 1981). States, throughout much of Mexico and Central America, to northwestern South America (western Colombia and northwest- Oryzomys Baird, 1857 ern Venezulela), including many continental- Mus, Harlan, 1837: 385 (part, description of shelf islands. palustris). CONTENTS: albiventer Merriam, 1901 (in- Arvicola, Bachman, 1854 [in Audubon and Bach- cluding molestus Elliot, 1903); antillarum man, 1846–1854]: 214 (part, description of Thomas, 1898; couesi Alston, 1876 (including oryzivora). apatelius Elliot, 1904; aquaticus J.A. Allen, Hesperomys (Oryzomys) Baird, 1857: 482. 1891; aztecus Merriam, 1901; azuerensis Bole, Oryzomys, Allen, 1890: 187 (first usage as genus). 1937; bulleri J.A. Allen, 1897; cozumelae Merriam, 1901; crinitus Merriam, 1901; ful- TYPE SPECIES: Mus palustris Harlan, 1837, gens Thomas, 1893; gatunensis Goldman, by original designation. 1912; goldmani Merriam, 1901; jalapae J.A. EMENDED DIAGNOSIS: A of medi- Allen and Chapman, 1897; lambi Burt, 1934; um-sized oryzomyine species (Cricetidae: mexicanus J.A. Allen, 1897; peragrus Mer- Sigmodontinae: ) characterized, riam, 1901; pinicola A. Murie, 1932; regillus in combination, by: supraorbital ridging well Goldman, 1915; richardsoni J.A. Allen, 1910; developed and interorbital shape cuneate; richmondi Merriam, 1901; rufinus Merriam, incisive foramina long, extending to or 1901; rufus Merriam, 1901; teapensis Mer- between the alveoli of the M1s; bony palate riam, 1901; zygomaticus Merriam 1901); long, projecting behind the posterior margins dimidiatus Thomas, 1905 ; gorgasi Hershko- of M3s, and posterolateral palatal pits vitz, 1971; nelsoni Merriam, 1898; palustris conspicuous and internally perforated; ecto- Harlan, 1837 (including argentatus Spitzer tympanic bulla medium-sized, revealing ven- and Lazell, 1978; coloratus Bangs, 1898; tral-medial wedge of the periotic; zygomatic floridanus Merriam, 1901; natator Chapman, plate broad, forming a moderately deep 1893; oryzivora Bachman, 1854 [in Audubon notch, and zygomatic arches sturdy, with and Bachman, 1854]; planirostris Hamilton, small but consistently formed jugal; interpa- 1955; sanibeli Hamilton, 1955; texensis J.A. rietal small and narrow, about half as wide as Allen, 1894); peninsulae Thomas, 1897. the caudal border of parietals; alisphenoid REMARKS: In the early taxonomic litera- strut absent; carotid arterial circulatory ture on North American forms (Merriam, pattern derived, lacking supraorbital and 1901; Goldman, 1918), Baird’s Oryzomys was infraorbital branches (sphenofrontal foramen broadly conceived and included species and squamosal-alisphenoid groove absent, groups and subgenera now removed to stapedial foramen absent or minute, posteri- separate genera (Gardner and Patton, 1976; or opening to alisphenoid canal irregularly Carleton and Musser, 1989; Weksler, 2003, formed or occluded). Molars pentalophodont 2006; Weksler et al., 2006). The phylogenet- (mesolophs/ids well developed) and cuspi- ically reified understanding of Oryzomys date, cusps bunodont with lingual and labial corresponds to the association of species series unevenly terraced; anterocone/id placed in the palustris-mexicanus or palustris broad, anteromedian flexus/id absent, ante- group, as designated, respectively, by Mer- roloph discrete and well formed; satellite riam (1901) and Goldman (1918). Thus, O. rootlets present on M1 (4-rooted) and m1 (3– albiventer, O. couesi mexicanus, O. nelsoni, 4-rooted). Eight mammae present (pectoral, and O. peninsulae share with O. palustris those postaxial, abdominal, and inguinal pairs); derived conditions that characterize the strict pedal digits 1 and 5 notably shorter than concept of Oryzomys (Carleton and Musser, central three; ungual tuft of hairs sparsely 1989; Carleton and Olson, 1999; Sa´nchez et developed, covering less than half of the nail; al., 2001) and that were amplified to infer the plantar surface squamate and pads small, phylogenetic position of the genus within the hypothenar pad minute or absent. oryzomyine radiation (Weksler, 2006). 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 117

The four western Mexican species further and 3–4 in his sample of O. palustris but exhibit the trenchant characteristics of the absent in that of O. couesi,andSa´nchez et couesi section within the O. palustris group, al. (2001) indicated that such webbing also as denoted by Goldman (1918: 20) based on characterized O. dimidiatus and O. gorgasi. several pelage and cranial traits. In the Based on examinations of the hind feet on palustris section, the pelage color is predom- dried skins and even fluid-preserved speci- inantly grayish brown, with long glistening mens of the taxa in western Mexico, we are guard hairs conspicuously overtopping the uncertain about the condition of this trait. body fur; the pelage color is generally Certainly, webbing is not so obviously reddish brown, with the guard hairs only developed as in O. palustris, but slight slightly longer than the body fur in the couesi webbing seemed to be present between digits section. Spacious, oval sphenopalatine vacu- 3–4 in some specimens. This character ities perforate the walls of the mesopterygoid definition and variation should be critically fossa, exposing much of the presphenoid reviewed within Oryzomys to better under- bone in specimens of O. palustris (e.g., see stand its taxonomic significance. character 36: state 0, as per Weksler, 2006); In summary, the palustris and couesi in forms of the couesi section, sphenopala- ‘‘sections’’ of Goldman’s palustris group tine vacuities are absent or weakly expressed are sufficiently characterized morphological- as short, narrow slits near the basisphenoid- ly to maintain them as the palustris and presphenoid suture (character 36: states 1 couesi species groups within the genus and 2, as per Weksler, 2006). Other differ- Oryzomys as now phylogenetically diag- ences reinforce Goldman’s perception of a nosed. For the character variation summa- division. Oryzomys couesi and the western rized above, the palustris group displays Mexican species all possess a small, circular more derived conditions and may possess in contrast to the more semiaquatic adaptations than the larger, ovate shape of the foramen as couesi group. Only O. palustris is currently observed in skulls of O. palustris. Schmidt recognized as a valid species within the and Engstrom (1994) drew attention to the palustris group; the principal alpha system- utility of this contrast in separating crania of atic uncertainty involves the status of O. palustris texensis and O. couesi aquaticus argentatus Spitzer and Lazell (1978), a form where their populations overlap along coast- described from the Keys that has al Texas, but the contrast holds between O. been treated as species or subspecies (see palustris and the four species in western Musser and Carleton, 2005, for review). The Mexico as well. We noticed that the two other seven species fit with the couesi group, sections also differ in the development of the but as underscored by our findings in anterolabial cingulum on the lower third western Mexico, comprehension of species molar: consistently present and forming a limits warrants additional research. distinct ledge in the couesi section; absent or represented as a short spur in members of Oryzomys albiventer Merriam, 1901 the palustris section (see Carleton and Musser, 1989: 44–45, for description and Oryzomys albiventer Merriam, 1901: 279. Type illustration of the condition in O. palustris). locality: Mexico, Jalisco, Ameca, 4000 ft; Plantar pads of the hindfoot are small in holotype: USNM 82236 (skin and skull), a male species of Oryzomys compared with many collected 6 February 1897 by E.W. Nelson and other oryzomyines, and the hypothenar pad E.A. Goldman. is typically absent in O. palustris or tiny if Oryzomys couesi albiventer, Goldman, 1918: 38 (name combination, allocated as valid subspe- present (Carleton and Musser, 1989; Weks- cies). ler, 2006). A hypothenar pad is consistently Oryzomys palustris albiventer, Hall, 1960: 173 present in all members of the couesi section (name combination, listed as valid subspecies). for which we could ascertain its condition, Oryzomys molestus Elliot, 1903: 145. Type locality: its size slightly smaller than interdigital pads Mexico, Jalisco, Ocotla´n, 5000 ft; holotype: 1 and 4. Weksler (2006) scored incipient FMNH 8667 (skin and skull), a male collected webbing as present between pedal digits 2–3 June 1901 by F.E. Lutz (not examined). 118 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Oryzomys couesi albiventer, Goldman, 1918: 38 OTU. Two specimens from localities in (molestus allocated as full synonym, invalid as northern Michoaca´n (Zamora and Quere´n- subspecies). daro), which Goldman (1918) referred to O. Oryzomys palustris albiventer, Hall, 1981: 609 c. regillus, are here considered examples of O. (molestus listed as full synonym, invalid as albiventer proper. subspecies). Although the distributional limits of O. EMENDED DIAGNOSIS: A species of the O. albiventer require further documentation, the couesi group characterized by a combination species so far appears to be endemic to the of large size (HFL < 34–38 mm), tail Mesa de Ana´huac, a small inland plateau relatively and absolutely long (TL < 140– (1200–1900 m) that is enclosed by southern 170 mm), skull heavily constructed (ONL < ranges of the Sierra Madre Occidental and by 31–34 mm), and molars correspondingly the Transverse Volcanic Range. Character- robust (CLM < 5.0–5.4 mm). ized by a subhumid climate and subtropical DISTRIBUTION: Central and eastern Ja- vegetation, the region corresponds to one of lisco, southern , and northern the five zoogeographic zones that Guerrero Michoaca´n at intermediate elevations, 1200– Va´zquez et al. (1995) identified within Jalisco. 1800 m (fig. 7). SPECIMENS EXAMINED: Mexico, Guana- REMARKS: Oryzomys albiventer is a large, juato: Salvatierra, 5775 ft (KU 66597). brightly colored species with a long tail. Jalisco: Ameca, 4000 ft (USNM 82235– Merriam’s (1901) key diagnostic feature, 82246); 6 km NE Contla, 1310 m (KU ‘‘underparts white,’’ and consequent choice 120611, 120612); 2 mi N and 0.5 mi W of the specific name are somewhat misleading. Guadalajara, 5200 ft (KU 31782, 31783); 27 The portion of ventral hairs is pale gray mi S and 12 mi W Guadalajara (KU 31784); and the distal half a pure white in the type La Barca (USNM 129038); 2 mi SE Mascota, series from Ameca. The white tips do not 5200 ft (KU 100509); Ocotla´n, 5000 ft obscure the basal gray and, to us, create a (USNM 120160–120163); 4 km W Tuxpan, general effect of pale gray to grayish-white 1380 m (KU 120613). Michoaca´n: Quere´n- underparts. In the specimens from Ocotla´n, daro (USNM 50180); Zamora, 4000 ft thegraybasesaremoreprominentandthe (USNM 120308). underparts appear medium gray. Other aspects of Merriam’s description are accurate, notably Oryzomys couesi mexicanus J.A. Allen, 1897 the large skull and long and wide molar rows which he emphasized in his comparisons with Oryzomys mexicanus J.A. Allen, 1897: 52. Type nearby species. Goldman (1918) repeated these locality: Mexico, Jalisco, Tonila, Hacienda San Marcos, 3500 ft; holotype: AMNH 2650/2128 salient contrasts in his rearrangement of (skin and skull), a male collected 30 December albiventer, mexicanus,andregillus as subspe- 1889 by Audley C. Buller. cies of O. couesi in western Mexico. Oryzomys couesi mexicanus, Goldman, 1918: 33 Goldman (1918) reallocated molestus El- (name combination, allocated as valid subspe- liot (1903) as a full synonym of Merriam’s cies). (1901) albiventer (then a subspecies of O. Oryzomys palustris mexicanus, Hall, 1960: 173 couesi), and we concur with their synonymy. (name combination, listed as valid subspecies). Elliot based his new species on a single Oryzomys bulleri J.A. Allen, 1897: 53. Type specimen, but four other individuals in locality: Mexico, Tepic [Nayarit], Valle de USNM from Ocotla´n, the type locality of Banderas; holotype: 8329/6663 (skin and skull), molestus, conform to his description. The a male collected 2 February 1893 by Audley C. type specimen of molestus is a strikingly large Buller. Oryzomys couesi mexicanus, Goldman, 1918: 33 individual (table 5), but it is, as noted by (bulleri allocated as full synonym, invalid as Goldman, an old adult and a male. Mea- subspecies). surements of the type fit within the size range Oryzomys palustris mexicanus, Hall, 1981: 610 of O. albiventer as we understand the species, (bulleri listed as full synonym, invalid as albeit at or toward the upper limits for most subspecies). variables; still, discriminant-function analysis Oryzomys rufus Merriam, 1901: 287. Type locality: decidedly assigned the type to the albiventer Mexico, Tepic [Nayarit], Santiago, 200 ft; 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 119

Fig. 10. Scatter plot of coronal length of molar rows versus length bony palate (in mm) for specimens of O. albiventer and O. couesi mexicanus from Jalisco. Values for type specimens of taxa described from Jalisco (albiventer Merriam, mexicanus Allen, molestus Elliot) and nearby Nayarit (bulleri Allen, rufus Merriam) are indicated.

holotype: USNM 91404 (skin and skull), a Morelos, southern , and northwestern female collected 20 June 1897 by E.W. Nelson Oaxaca (fig. 7); most localities from sea level and E.A. Goldman. to 1000 m, exceptional occurrence at 1525 m Oryzomys couesi mexicanus, Goldman, 1918: 33 in the Sierra de Autla´n, Jalisco. (rufus allocated as full synonym, invalid as subspecies). REMARKS: J.A. Allen (1897) named mex- Oryzomys palustris mexicanus, Hall, 1981: 610 icanus and bulleri as new species based on one (rufus listed as full synonym, invalid as subspe- specimen of the former and two of the latter. cies). Although their type localities are relatively Oryzomys couesi lambi Burt, 1934: 107. Type near one another in western Mexico, he locality: Mexico, Sonora, San Jose´ de Guaymas; differentiated mexicanus with respect to O. holotype: California Academy of Sciences 51024 palustris and bulleri to O. couesi rather than (skin and skull), a female collected 20 January to one another. Working with larger series 1933 by C.C. Lamb (not examined). and better geographic representation, Gold- Oryzomys palustris lambi, Hall, 1960: 173 (name man (1918) recognized the fundamental combination, listed as valid subspecies). similarity of Allen’s taxa and adopted mex- DISTRIBUTION: Coastal plain and adjoin- icanus as the subspecific epithet, with bulleri ing lower Pacific-facing slopes from central as full synonym, for the populations of Sonora to southeastern Oaxaca; interior smaller Oryzomys widely distributed along river valleys to central Michoaca´n, southern Pacific coastal plain. Although the damaged 120 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331 skull of the type of mexicanus prevented its conforming to his Arid Lower Tropical Zone use in our morphometric analyses, bivariate (and Goldman, 1951) and essentially repeated scatterplots of measurable variables depict in later taxonomic compendia of mammals both mexicanus and bulleri as falling within (Hall and Kelson, 1959; Hall, 1981). Howev- the range of samples drawn from coastal er, specimens from certain localities that plain (fig. 10), in contrast to those originating Goldman allocated to aztecus also represent from the inland plateau region (i.e., albiventer O. couesi mexicanus proper in our judgment, and molestus). and these reallocations extend its range into Merriam (1901: 287) considered his new interior Mexico along the valleys of the Rı´os species rufus, based on a single specimen, to Tepalcatepec and Balsas. A geographic range be ‘‘similar in general to bulleri but smaller that covers coastal plain, Pacific-facing lower and redder.’’ The type specimen is a small slopes of western sierras, and the Balsas individual (table 5) but falls within the Depression resembles other cricetid variational sphere of the larger series from found in western Mexico (e.g., Sigmodon Nayarit and Jalisco, whether viewed from a mascotensis, Hodomys alleni, Peromyscus univariate or multivariate perspective (figs. 6, perfulvus, Osgoodomys banderanus) and pre- 10). The contrasts that impressed Merriam sumably reflects a shared biogeographic can be largely attributed to the age and gender history. The pattern is nearly exactly congru- differences of the types of bulleri (young adult ent to the biotic zone distilled from the male) and rufus (old adult female) and the distributions of the Mexican herpetofauna tendency of the pelage of aged Oryzomys to (Flores-Villela, 1993) but is not so precisely become more reddish. Again, we agree with mirrored in those recognized in biogeographic Goldman’s (1918) synonymy of rufus under syntheses of birds and mammals (Escalante O. couesi mexicanus. Pliego et al., 1993; Fa and Morales, 1993). The three topotypes of lambi Burt (1934) at SPECIMENS EXAMINED: Mexico, Colima: hand generally resemble other examples of the Armerı´a, 200 ft (USNM 33284/45312–33285/ Oryzomys species that inhabits coastal low- 45313, 33292/45320–33297/45325); 4 km S lands in western Mexico (e.g., fig. 6, table 5), and 2 km W El Chavarı´n (CM 79555, 79556); here called O. couesi mexicanus, and we Hacienda Magdalena, 1500 ft (USNM 33299/ include it among the synonyms of this 45327, 33300/45328); 5 mi N Manzanillo (KU ‘‘subspecies.’’ Our maintenance of a trinomial 87515–87534); Paso del Rı´o, 200 ft (UMMZ is pro forma in this instance because we 100656–100669, 100689); 3 mi N Santiago cannot here resolve the larger issue of the (KU 87490–87492, 87653); 1 mi NE Santiago, taxonomic and distributional limits of O. 10 ft (KU 36867–36873); 2 km E Santiago couesi sensu stricto. The subspecific range of (KU 87498); 3 km E Santiago (KU 87499– lambi is depicted as restricted to its type 87514); 1 mi S and 4 mi W Santiago, 10 ft locality in west-central Sonora, whether as a (KU 36874); 12 km W Santiago (KU 87493– race of O. couesi (Hall and Kelson, 1959; 87497). Guerrero: 5 km N Agua del Obispo, Alvarez-Castan˜eda and Corte´s-Calva, 1999) 3250 ft (KU 99899); 5.5 km N Agua del or O. palustris (Hall, 1981). We note that Obispo, 3400 ft (KU 99898); near Ometepec, certain specimens from northern Sinaloa (e.g., 200 ft (USNM 71400–71405, 71419); Zihua- 2 mi E Costa Rica, 3 mi N Higueras) are tanejo Bay (UMMZ 80926, 80927, 80929, larger than those from southern Sinaloa and 80931); Petacalco, sea level (UMMZ 113850– farther south along coastal plain in Nayarit, 113852); Tlalixtaquilla, 4200 ft (USNM Jalisco, and Colima, and may belong with 70277); 2 km SSE Zihuatanejo, Las Gatas Burt’s lambi. Denser sampling in central (KU 35371–35379). Jalisco: 11 mi SW Autla´n, Sinaloa and refined analyses are needed to 2000 ft (KU 105742); Sierra de Autla´n, 5000 ft determine the relationships of these popula- (UMMZ 95476, 95477); Cihuatla´n, 15 ft (KU tions and whether lambi deserves formal 109262–109265); 15 KM NW Cihuatla´n subspecific retention. (87454); Cuitzamala, 25 ft (KU 105743– Goldman (1918) viewed the distribution of 105763); 2 km NW Emiliano Zapata, 20 m the subspecies mexicanus to adhere to Pacific (KU 120615); Jilotla´n de Los Dolores, 2400 ft coastal plain and contiguous foothills, a range (KU 111830, 111831); 8 mi E Jilotla´n de Los 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 121

Dolores, 2000 ft (KU 111832–111837); 6 mi E 553966, 554089); San Blas, 75 ft (USNM Limo´n, 2700 ft (KU 103799); 2 mi S La 89206–89208, 89240); 0.5 E San Blas, 10 ft Cuesta, 2400 ft (KU 111829); 2 mi N La (KU 36852–36859); 1 mi S San Francisco, 50 Resolana, 1500 ft (UMMZ 95478, 95479); 2 ft (KU 64490–64493); 2 mi SW San Juan de mi SW La Resolana, 1100 ft (KU 40125, Abajo (USNM 512558–512561, 512759, 40126–40129, 40138, 40139); 1 mi N Barra de 512760); 2 mi E San Pedro Lagunillas, E side Navidad (KU 97242–97244, 97246–97248); 3 lake, 1300 m (USNM 509474–509493, mi NW Barra de Navidad, 30 ft (KU 97239– 510564–510567); 15 km S Santa Cruz, Playa 97241); 5 km NNW Barra de Navidad (KU Colorado (USNM 553963–553965, 554088); 87455–87465, 87467–87472, 87474–87489); Santiago, 200 ft (USNM 91403, 91404); 0.5 mi N Navidad, 25 ft (UMMZ 100654, Teponahuaxtla (USNM 509451–509454, 100655, 101268–101270); 3 mi E Navidad, 50 510561); 2 mi WNW Valle de Banderas (KU ft (UMMZ 100641–100653); Navidad Bay 64483–64489). Oaxaca: Cuicatla´n, 1800–1900 (UMMZ 80912–80914, 80924); 14.5 mi S ft (USNM 69654, 69660); 1 km N Cuicatla´n, Pihuamo, 1100 ft (KU 111820–111827); 4 mi 560 m (KU 29490–29492); 3 km NNE NNE Puerto Vallarta, 50 ft (KU 64494– Cuicatla´n, 600 m (KU 29487–29489); 1 km 64498); 4 mi SW Puerto Vallarta, 20 ft (KU S Cuicatla´n, 590 m (KU 29493); 2 km SSW 111828); 20 km WNW Purificacio´n, 1400 ft Cuicatla´n, 570 m (KU 29494); Huilotepec, (KU 93021, 93652); 30 km N and 10 km E 100 ft (USNM 73408, 73726); Llano Grande, Santiago (KU 87489); El Tabaco, 200 ft (KU 200 ft (USNM 71406, 71407, 71921, 73727, 107669–107674); Tenacatita Bay (UMMZ 73730); Plumas, 3000 ft (USNM 71415, 71417, 80897, 80906, 80911). Michoaca´n:16miS 71418); Puerto Angel, 100–200 ft (USNM Arteaga, 800 ft (KU 40132–40137, 40140); 0.5 71408–71414, 73745); Tehuantepec (USNM mi SE Coalcoma´n, 3600 ft (UMMZ 100670– 73407). Puebla: Piaxtla, 3900 ft (USNM 100672); near La Huacana (USNM 126497– 70273–70276). Sinaloa: 2 mi E Costa Rica, 126501); 1 mi E La Mira, 300 ft (KU 40141– 100 ft (KU 100502–100508); 24 km S Gua- 40144); 9 mi S Lombardı´a, 1500 ft (KU save, 20 ft (KU 90316); 3 mi N Higueras, 30 ft 38652); 7 mi S Tumbiscatio, 2700 ft (KU (KU 96669–96674); 6 mi NNW Teacpan, Isla 40130); Parque Nacional Uruapan (UMMZ Palmito del Verde (KU 98960, 98961); 6 mi W 110560, 110561). Morelos: Puente de Ixtle La Concha, 10 ft (KU 98962, 99156); 5 mi (USNM 126943, 127496, 127497). Nayarit: NW Mazatla´n (KU 85825–85829); Rosario, Arroyo de Jiguite, Rı´o Santiago, 100 m 100 ft (USNM 91405); 8 km N Villa Unio´n, (USNM 523637, 523638); Aticama, 10 ft 450 ft (KU 96099, 96100); 1 mi N Zaragoza, (KU 36860–36864); Chacala, 30 m (USNM 30 ft (KU 90314, 90315). Sonora: San Jose´de 509467–509473, 523631–523636); 1 mi S Guaymas (UCLA 51021, 51026, 51038). Cuautla, near sea level (USNM 509389– 509450, 510551–510560); 1.2 mi S (by road) Oryzomys nelsoni Merriam, 1898 El Casco, Rı´o Chilte, 480 ft (USNM 511645– 511652, 511751, 511752, 511884–511889); El Oryzomys nelsoni Merriam, 1898: 15. Type locality: Refilion, 850 m (USNM 509463–509466); 3.5 Mexico, Nayarit, Tres Marı´as Islands, Marı´a mi E (by road) El Venado, 100 m (USNM Madre Island, 800 ft; holotype: USNM 89200 509455, 510562); Estanzuela, 1380 m (USNM (skin and skull), an adult male collected 13 May 1897 by E.W. Nelson and E.A. Goldman. 509474); 1 mi E Ixtlan, 4000 ft (UMMZ Oryzomys palustris nelsoni, Hershkovitz, 1971: 704 95475); 2 mi E Jalcocotan, 500 m (USNM (name combination, listed as valid subspecies). 523624–523630, 524358, 524359); 1 mi NE Las Piedras, Rı´o Tigrera (USNM 523639, DISTRIBUTION: Known only from Isla 523640); 5 mi S Las Varas, 150 ft (KU 40124); Marı´a Madre, the largest of the Tres Marı´as, 8 mi SSW Las Varas (KU 64480–64482); 4 mi Nayarit. NE San Blas, 100 ft (UMMZ 103184– REMARKS: Merriam (1898, 1901) aptly 103187); 8.8 mi E San Blas, Paso de Soquilpa, captured the distinguishing characteristics of 100 m (USNM 509456–509462, 510563); 10 his new form, in particular its large skull, mi E Ruiz, Platanares (KU 85830, 85831, massive and strongly decurved rostrum, and 85837); 4 mi E Playa Los Corchos (USNM large, broad incisors (figs. 8, 9). Dimensions 122 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

that reflect Merriam’s qualitative descrip- REMARKS: Merriam (1901) and Goldman tions, particularly those measured on the (1918) were clearly impressed by the differenti- rostrum, figure prominently among the ating traits of O. peninsulae, as emphasized by several variables that account for isolation Thomas (1897), and many of their qualitative of the four specimens of O. nelsoni in characterizations of its distinctly shaped skull discriminant-function analyses (table 4). contributed to its morphometric separation Merriam also mentioned the island form’s (e.g., ‘‘high braincase,’’ ‘‘large and broadly large body size, long tail, and big hind feet; in open palatal slits,’’ ‘‘rather large molars’’). The relative and absolute size, nelsoni possesses pelage contrast between the grayish-brown the longest tail among the species of the O. forequarters and fulvous brown hindquarters couesi group in western Mexico (table 2). is still appreciable in the century-old, now Examples of O. albiventer lack such a robust slightly faded skins that Nelson and Goldman rostrum and upper incisors as found in O. collected at San Jose´ del Cabo and is unmatched nelsoni but possess heavier molars (table 2). for its constancy in other western Mexican We must accept that the morphology evident Oryzomys. The orthodont shape of the upper in the small museum series of the Tres Marı´as incisors (fig. 9) impresses us as also diagnostic Oryzomys is typical of the species. Sadly, O. of the species, but the feature was not nelsoni represents an instance where the mentioned by the aforementioned authors. investigator’s sample statistics and popula- The Oryzomys population isolated at the tion parameters are one and the same. southernmost tip of the Baja Peninsula was SPECIMENS EXAMINED: Mexico, Nayarit: maintained as a valid species until grandiosely Tres Marı´as Islands, Isla Marı´a Madre synonymized under O. palustris by Hershko- (USNM 89200–89203). vitz (1971). Hall (1981: 610–611) later pro- vided some post hoc rationale for its subspe- Oryzomys peninsulae Thomas, 1897 cific rank within O. palustris and considered ‘‘the degree of difference between O. peninsu- Oryzomys peninsulae Thomas, 1897: 548. Type lae and O. palustris mexicanus of the eastern locality: Mexico, Lower California [Baja Cali- fornia Sur], Santa Anita; holotype: BMNH coast of the Gulf of California to be less than 98.3.1.107 (skin and skull), a male collected on between some intergrading subspecies of O. 23 May 1896 by Dane Coolidge. palustris (e.g., O. palustris texensis and O. p. Oryzomys palustris peninsulae, Hershkovitz, 1971: aquaticus).’’ Hall’s precept for the synonymy 704 (name combination, listed as valid subspe- of peninsulae, the purported intergradation of cies). texensis and aquaticus, has proven to be Oryzomys couesi peninsularis, Alvarez-Castan˜eda, erroneous, the latter forms instead represent- 1994: 99 (incorrect subsequent spelling, assess- ing nonintrogressing races of the species O. ment of conservation status). palustris and O. couesi, respectively (Benson Oryzomys couesi peninsulae, Alvarez-Castan˜eda and Gehlbach, 1979; Schmidt and Engstrom, and Corte´s-Calva, 1999: 448 (name combina- 1994). As we document above, the morpho- tion, treated as valid subspecies). logical identity of O. peninsulae is sharply and EMENDED DIAGNOSIS: A species of the O. uniquely circumscribed relative to popula- couesi group characterized by a combination tions of O. couesi mexicanus and other of moderate size (larger than O. couesi Oryzomys in western Mexico. Such a ‘‘degree mexicanus, smaller than O. albiventer)and of difference’’ clearly advises specific recog- dorsal pelage of head and forequarters nition until controverted by other evidence. predominantly gray toned. Cranium with SPECIMENS EXAMINED: Mexico, Baja Cal- deep braincase, zygomatic arches broadly ifornia Sur: Santa Anita (BMNH 98.3.1.107, spreading and squarely shaped, and incisive USNM 93398); San Jose´ del Cabo (USNM foramina long and exceptionally broad. 146607–146620, 147340). Upper incisors relatively deep and curvature more nearly orthodont. ACKNOWLEDGMENTS DISTRIBUTION: Known only from two localities in southernmost Baja California Foremost, we appreciate the access to Sur, Nayarit. specimens, through loans or museum visits, 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 123 as facilitated by the following museum staff: the American Museum of Natural History 9 Guy G. Musser (AMNH), Paula Jenkins and (13): 197–208. Daphne Hill (BMNH), Sue McLaren (CM), Alston, E.R. 1876 [1877]. On two new species of William T. Stanley (FMNH), Thor Holmes Hesperomys. Proceedings of the Zoological and Bob Timm (KU), Fritz Hertel (UCLA), Society of London 1876: 755–757. Alvarez-Castan˜eda, S.T. 1996. Los mamı´feros del and Steve Hinshaw and Phil Myers estado de Morelos., Centro de Investigaciones (UMMZ). Bill Stanley (FMNH) graciously Biolo´gicas del Noroeste: La Paz, Me´xico, accommodated our request to provide mea- 211 pp. surements of the type specimen of Oryzomys Alvarez-Castan˜eda, S.T. 1994. Current status of molestus. John Steiner (Senior Photographer, the rice rat, Oryzomys couesi peninsularis. NMNH Branch, Smithsonian Photographic Southwestern Naturalist 39: 99–100. Services) photographed the skulls illustrated Alvarez-Castan˜eda, S.T., and P. Corte´s-Calva. in figures 8 and 9, and Dan Cole (Office of 1999. Familia Muridae. In S.T. Alvarez-Casta- Information Technology, NMNH) offered n˜eda and J.L. Patton (editors), Mamı´feros del assistance and advice on GIS mapping. Data noroeste de Me´xico: 445–568, Centro de In- vestigaciones Biolo´gicas del Noroeste: La Paz, collection for this project was initiated when Me´xico, 583 pp. the junior author was on sabbatical to the Alvarez-Castan˜eda, S.T., and L. Me´ndez. 2003. USNM Division of Mammals (1998–1999), Oryzomys nelsoni. Mammalian Species 735: in a postdoctoral position provided by Don 1–2. E. Wilson and funded through the Smithso- Audubon, J.J., and J. Bachman. 1846 [1845]–1854. nian Institution’s Office of Fellowships and The viviparous quadrupeds of North America. Grants. Duke Rogers and Marcelo Weksler Privately published by J.J. Audubon, New reviewed the manuscript and offered com- York 1: 1–389 [1846], 2: 1–334 [1851], 3: 1–348 ments that improved it final contents. [1854]. Baird, S.F. 1857. Mammals: General report upon the zoology of the several Pacific railroad REFERENCES routes. Vol. 8, part 1. In Reports of explorations and surveys to ascertain the most practicable Allen, J.A. 1890. Notes on collections of mammals and economical route for a railroad from the made in Central America and southern Mexico, Mississippi River to the Pacific Ocean Senate by Dr. Audley C. Buller, with descriptions of Executive Document 78: Washington, DC, new species of the genera Vespertilio, Sciurus, 757 pp. and Lepus. Bulletin of the American Museum of Bangs, O. 1898. The land mammals of peninsular Natural History 3 (11): 175–194. Florida and the coast region of Georgia. Allen, J.A. 1891. Notes on new or little-known Proceedings of the Boston Society of Natural North American mammals, based on recent History 28: 157–235. additions to the collection of mammas in the Benson, D.L., and F.R. Gehlbach. 1979. Ecolog- American Museum of Natural History. Bulletin ical and taxonomic notes on the rice rat of the American Museum of Natural History 3 (Oryzomys couesi) in Texas. Journal of Mam- (20): 263–310. malogy 60: 225–228. Allen, J.A. 1894. On the mammals of Aransas Bole, B.J., Jr. 1937. Annotated list of the mammals County, Texas, with descriptions of new forms of the Mariato River District of the Azuero of Lepus and Oryzomys. Bulletin of the Amer- Peninsula. In J.W. Aldrich and B.P. Bole, Jr ican Museum of Natural History 6 (6): 165– (editors), The birds and mammals of the western 198. slope of the Azuero Peninsula (Republic of Allen, J.A. 1897. Further notes on mammals Panama): 140–196. Scientific Publications of the collected in Mexico by Dr. Audley C. Buller, Cleveland Museum of Natural History 7: 1–196. with descriptions of new species. Bulletin of the Bradley, R.D., D.J. Schmidly, and C.W. Kilpa- American Museum of Natural History 9 (5): trick. 1996. The relationships of Peromyscus 47–57. sagax to the P. boylii and P. truei species groups Allen, J.A. 1910. Additional mammals from in Mexico based on morphometric, karyotypic, Nicaragua. Bulletin of the American Museum and allozymic data. In H.H. Genoways and R.J. of Natural History 28 (9): 87–115. Baker (editors), Contributions in mammalogy: a Allen, J.A., and F.M. Chapman. 1897. On a memorial volume in honor of J Knox Jones, Jr.: collection of mammas from Jalapa and Las 95–106. Lubbock, TX: Museum of Texas Tech Vigas, state of Vera Cruz, Mexico. Bulletin of University, il + 315 pp. 124 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Bradley, R.D., D.J. Schmidly, and R.D. Owen. Peromyscus boylii ambiguus based on morpho- 1989. Variation in the glans penes and bacula logic and molecular data. Journal of - among Latin American populations of the ogy 80: 1214–1231. Peromyscus boylii species complex. Journal of Ceballos, G., J. Arroyo-Cabrales, and R.A. Me- Mammalogy 70: 712–725. dellı´n. 2002. The mammals of Me´xico: compo- Burt, W.H. 1934. A new rice rat (Oryzomys) from sition, distribution, and conservation status. Sonora, Mexico. Proceedings of the Biological Occasional Papers Museum of Texas Tech Society of Washington 47: 107–108. University 218: 1–27. Carleton, M.D. 1977. Interrelationships of popu- Chapman, F.M. 1893. Description of a new lations of the Peromyscus boylii species group subspecies of Oryzomys from the Gulf states. (Rodentia, Muridae) in western Mexico. Occa- Bulletin of the American Museum of Natural sional Papers of the Museum of Zoology History 5 (4): 43–46. University of Michigan 675: 1–47. Cody, M.L., R. Moran, J. Rebman, and H. Carleton, M.D. 1979. Taxonomic status and Thompson. 2002. Plants. In T.J. Case, M.L. relationships of Peromyscus boylii from El Cody, and E. Ezcurra (editors), A new island Salvador. Journal of Mammalogy 60: 280–296. of the Sea of Cortez: 63–111. Carleton, M.D., R.D. Fisher, and A.L. Gardner. Oxford: Oxford University Press. 1999. Identification and distribution of cotton Cody, M.L., and E. Velarde. 2002. Land birds. In rats, genus Sigmodon (Muridae: Sigmodonti- T.J. Case, M.L. Cody, and E. Ezcurra (editors), nae), of Nayarit, Mexico. Proceedings of the A new island biogeography of the Sea of Cortez: Biological Society of Washington 112: 813–856. 271–312. Oxford: Oxford University Press. Carleton, M.D., and G.G. Musser. 1989. System- Corbet, G.B., and J.E. Hill. 1991. A world list of atic studies of oryzomyine rodents (Muridae, mammalian species. 3rd ed. London: British Sigmodontinae): a synopsis of Microryzomys. Museum (Natural History), 243 pp. Bulletin of the American Museum of Natural Elliot, D.G. 1903. A list of a collection of Mexican History 191: 1–83. mammals with descriptions of some apparently Carleton, M.D., and G.G. Musser. 1995. System- new forms. Field Columbian Museum Publica- atic studies of oryzomyine rodents (Muridae: tion 71, Zoological Series 3: 141–149. Sigmodontinae): definition and distribution of Elliot, D.G. 1904. Descriptions of apparently new (Bangs, 1902). Proceed- species and subspecies of mammals and a new ings of the Biological Society of Washington 108 generic name proposed. Field Columbian Mu- (2): 338–369. seum Publication 90, Zoological Series 3: Carleton, M.D., and S.L. Olson. 1999. Amerigo 263–270. Vespucci and the rat of Fernando de Noronha: Engstrom, M.D. 1984. Chromosomal, genic, a new genus and species of Rodentia (Muridae: and morphological variation in the Oryzomys Sigmodontinae) from a volcanic island off melanotis species group. Unpublished PhD. Brazil’s continental shelf. American Museum dissertation Texas A & M University, 171 pp. Novitates 3256: 1–59. Escalante Pliego, P., A.G. Navarro-Siguenza, and Carleton, M.D., D.E. Wilson, A.L. Gardner, and A.T. Peterson. 1993. A geographic, historical, M.A. Bogan. 1982. Distribution and systematics and ecological analysis of avian diversity in of Peromyscus (Mammalia: Rodentia) of Mexico. In T.P. Ramamoorthy, R. Bye, A. Lot, Nayarit, Mexico. Smithsonian Contributions and J. Fa (editors), Biological diversity of to Zoology 352: 1–46. Mexico: origins and distribution: 281–307. Carren˜o, A.L., and J. Helenes. 2002. Geology and New York: Oxford University Press, xxxix + ages of the islands. In T.J. Case, M.L. Cody, 812 pp. and E. Ezcurra (editors), A new island bioge- Esher, R.J., J.L. Wolfe, and J.N. Layne. 1978. ography of the Sea of Corte´s: 14–40. Oxford: Swimming behavior of rice rats and cotton rats. Oxford University Press, xvii + 669 pp. Journal of Mammalogy 59: 551–558. Carroll, D.S., L.L. Peppers, and R.D. Bradley. Fa, J.A., and L.M. Morales. 1993. Patterns of 2005. Molecular systematics and phylogeogra- mammalian diversity in Mexico. In T.P. Rama- phy of the Sigmodon hispidus species group. In moorthy, R. Bye, A. Lot, and J. Fa (editors), V. Sa´nchez-Cordero and R.A. Medellı´n (edi- Biological diversity of Mexico: origins and tors), Contribuciones mastozoolo´gicas en home- distribution: 319–361. New York: Oxford Uni- naje a Bernardo Villa, 87–100. Me´xico: Instituto versity Press, xxxix + 812 pp. de Biologı´a e Instituto de Ecologı´a, UNAM: Flores-Villela, O. 1993. Herpetofauna of Mexico: CONABIO. distribution and endemism. In T.P. Rama- Castro-Campillo, A., H.R. Roberts, D.J. Schmidly, moorthy, R. Bye, A. Lot, and J. Fa (editors), and R.D. Bradley. 1999. Systematic status of Biological diversity of Mexico: origins and 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 125

distribution: 253–280. New York: Oxford Uni- the United States, Mus palustris. American versity Press, xxxix + 812 pp. Journal of Science 31: 385–386. Forys, E.A., and R.D. Dueser. 1993. Inter- Hershkovitz, P. 1971. A new rice rat of the island movements of rice rats (Oryzomys Oryzomys palustris group (Cricetinae, Muridae) palustris). American Midland Naturalist 130: from northwestern Colombia, with remarks on 408–412. distribution. Journal of Mammalogy 52: Gardner, A.L., and J.L. Patton. 1976. Karyotypic 700–709. variation in oryzomyine rodents (Cricetinae) Honacki, J.H., K.E. Kinman, and J.W. Koeppl. with comments on chromosomal evolution in 1982. Mammal species of the world, a taxo- the Neotropical cricetine complex. Occasional nomic and geographic reference. Lawrence, KS: Papers of the Museum of Zoology Louisiana Allen Press, 694 pp. State University 49: 1–48. Humphrey, S.R., and H.W. Setzer. 1989. Geo- Genoways, H.H., and J.K. Jones, Jr. 1971. Second graphic variation and taxonomic revision of rice specimen of Oryzomys dimidiatus. Journal of rats (Oryzomys palustris and O. argentatus)of Mammalogy 52: 833–834. the United States. Journal of Mammalogy 70: Goldman, E.A. 1912. Descriptions of twelve new 557–570. species and subspecies of mammals from IUCN, The World Conservation Union. 2006. The Panama. Smithsonian Miscellaneous Collec- IUCN red list of threatened species (www. tions 56 (36): 1–11. iucnredlist.org; accessed May 2007). Goldman, E.A. 1915. Five new rice rats of the Jones, J.K., Jr., and M.D. Engstrom. 1986. genus Oryzomys from Middle America. Pro- Synopsis of the rice rats (Genus Oryzomys)of ceedings of the Biological Society of Washing- Nicaragua. Occasional Papers of the Museum ton 27: 127–130. Texas Tech University 103: 1–23. Goldman, E.A. 1918. The rice rats of North Jones, J.K., Jr., and T.E. Lawlor. 1965. Mammals America (Genus Oryzomys). North American from Isla Cozumel, Mexico, with description of Fauna 43: 1–100. a new species of harvest mouse. University of Kansas Publications of the Museum of Natural Goldman, E.A. 1951. Biological investigations in History 16: 409–419. Mexico. Smithsonian Miscellaneous Collections Lawlor, T.E. 1983. The mammals. In T.J. Case 115: 1–476. and M.L. Cody (editors), Island biogeography Guerrero Va´zquez, S., J. Te´llez Lopez, and R. in the Sea of Cortez: 265–289 and 482–496 Ampara´n Salido. 1995. Los mamı´feros de (appendix 10). Berkeley: University of Califor- Jalisco: ana´lisis zoogeogra´fico. BIOTAM 6: nia Press, 508 pp. 13–30. McFarlane, D.A., and A.O. Debrot. 2001. A new Hall, E.R. 1960. Oryzomys couesi only subspecifi- species of extinct oryzomyine rodent from the cally different from the , Oryz- Quaternary of Curac¸ao, Netherlands Antilles. omys palustris. Southwestern Naturalist 5: Caribbean Journal of Science 37: 182–184. 171–173. Medellı´n, X.L., and R.A. Medellı´n. 2006. Oryz- Hall, E.R. 1981. The mammals of North America. omys couesi (Alston, 1877). In G. Ceballos and 2nd ed., Vol. 2: 601–1181. New York: John G. Oliva (editors), Los mamı´feros silvestres de Wiley. Me´xico: 709–710. Me´xico: Comisio´n Nacional Hall, E.R., and K.R. Kelson. 1959. The mammals para el Conocimiento y Uso de la Biodiversidad of North America., Vol. 2: 547–1083. New and Fondo de Cultura Econo´mica, 986 pp. York: Ronald Press. Merriam, C.H. 1898. Mammals of the Tres Marias Hamilton, W.J., Jr. 1946. Habits of the swamp rice Islands, off western Mexico. Proceedings of the rat, Oryzomys palustris palustris (Harlan). Biological Society of Washington 12: 13–19. American Midland Naturalist 36: 730–736. Merriam, C.H. 1901. Synopsis of the rice rats Hamilton, W.J., Jr. 1955. Two new rice rats (genus Oryzomys) of the United States and (Genus Oryzomys) from Florida. Proceedings Mexico. Proceedings of the Washington Acad- of the Biological Society of Washington 68: emy of Sciences 3: 273–295. 83–96. Miller, G.S., and R. Kellogg. 1955. List of North Handley, C.O., Jr. 1966. Checklist of the mammals American Recent mammals. Bulletin of the of Panama. In R.L. Wenzel and V.J. Tipton United States National Museum 205: xii, 1–954. (editors), Ectoparasites of Panama: 753–795. Morgan, G.S. 1993. Quaternary land vertebrates Chicago: Field Museum of Natural History, xii of Jamaica. In R.M. Wright and E. Robinson + 861 pp. (editors), Biostratigraphy of Jamaica. Geologi- Harlan, R. 1837. Description of a new species of cal Society of America Memoir 182: 417–442. quadruped, of the order Rodentia, inhabiting Boulder, CO: Geological Society of America. 126 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 331

Murie, A. 1932. A new Oryzomys from a pine ridge the mitochondrial cytochrome b gene. Journal in British Honduras. Occasional Papers of the of Mammalogy 83: 396–407. Museum of Zoology University of Michigan Rennert, P.D., and C.W. Kilpatrick. 1987. Bio- 245: 1–3. chemical systematics of Peromyscus boylii.II. Musser, G.G., and M.D. Carleton. 1993. Family Chromosomally variable populations from east- Muridae. In D.E. Wilson and D.M. Reeder ern and southern Mexico. Journal of Mammal- (editors), Mammal species of the world, a ogy 68: 799–811. taxonomic and geographic reference. 2nd ed.: Sa´nchez, H.J., J. Ochoa G., and R.S. Voss. 2001. 501–755. Washington DC: Smithsonian Institu- Rediscovery of (Rodentia: tion Press, 1206 pp. Muridae), with notes on and natural Musser, G.G., and M.D. Carleton. 2005. Super- history. Mammalia 65: 205–214. family Muroidea. In D.E. Wilson and D.M. Schmidly, D.J., R.D. Bradley, and P.S. Cato. Reeder (editors), Mammal species of the 1988. Morphometric differentiation and tax- world, a taxonomic and geographic reference. onomy of three chromosomally characterized 3rd ed.: 894–1531. Baltimore, MD: Johns groups of Peromyscus boylii from east-central Hopkins University Press, Vol. 2, xvii + Mexico. Journal of Mammalogy 69: 462–480. 745–2142. Schmidt, C.A., and M.D. Engstrom. 1994. Genic Musser, G.G., M.D. Carleton, E.M. Brothers, and variation and systematics of rice rats (Oryzomys A.L. Gardner. 1998. Systematic studies of palustris species group) in southern Texas and oryzomyine rodents (Muridae: Sigmodontinae): northeastern , Mexico. Journal of diagnoses and distributions of species formerly Mammalogy 75: 914–928. assigned to Oryzomys ‘‘capito.’’ Bulletin of the Sharp, H.F., and Jr. 1967. Food ecology of the American Museum of Natural History 236: rice rat, Oryzomys palustris (Harlan), in a 1–376. Georgia salt marsh. Journal of Mammalogy Myers, P., and M.D. Carleton. 1981. The species 48: 557–563. of Oryzomys (Oligoryzomys) in Paraguay and Spitzer, N.C., and J.D. Lazell. 1978. A new rice rat the identity of Azara’s ‘‘Rat sixie`me ou rat a` (genus Oryzomys) from Florida’s Lower Keys. tarse noir.’’ Miscellaneous Publications Muse- Journal of Mammalogy 59: 787–792. um of Zoology, University of Michigan 161: Svihla, A. 1931. Life history of the Texas rice rat 1–41. (Oryzomys palustris texensis). Journal of Mam- Nelson, E.W. 1899a. General description of the malogy 12: 238–242. Tres Marias Islands, Mexico. North American Fauna 14: 7–13. SYSTAT Software Inc, SYSTAT for Windows, Nelson, E.W. 1899b. Mammals of the Tres Marias Version 10.2.01. Chicago, Illinois. Islands. North American Fauna 14: 14–19. Thomas, O. 1893. Notes on some Mexican Nelson, E.W. 1922. Lower California and its Oryzomys. Annals and Magazine of Natural natural resources. Memoirs of the National History, series 6, 11: 402–405. Academy of Sciences 16: 1–194. Thomas, O. 1897. Description of new bats and Osgood, W.H. 1909. Revision of the mice of the rodents from America. Annals and Magazine of American genus Peromyscus. North American Natural History, series 6, 20: 544–553. Fauna 28: 1–285. Thomas, O. 1898. On indigenous Muridae in Paradiso, J.L. 1960. Size variation in the rice rat. the West Indies; with the description of a new Journal of Mammalogy 41: 516–517. Mexican Oryzomys. Annals and Magazine of Park, A.W., and B.J.A. Nowosielski-Slepowron. Natural History, series 7, 1: 176–180. 1972. Biology of the rice rat (Oryzomys Thomas, O. 1905. New neotropical Molossus, palustris natator) in a laboratory environment. , Proechimys,andAgouti, with a note Zeitschrift fur Saugetierkunde 37: 42–51. on the genus Mesomys. Annals and Magazine of Park, A.W., M. Benstead, and B.J.A. Nowosielski- Natural History, series 7, 15: 584–591. Slepowron. 1974. Biology of the rice rat Tiemann-Boege, I., C.W. Kilpatrick, D.J. (Oryzomys palustris natator) in a laboratory Schmidly, and R.D. Bradley. 2000. Molecular environment II. General skull morphology. of the Peromyscus boylii species Zeitschrift fur Saugetierkunde 39: 103–109. group (Rodentia: Muridae) based on mito- Peppers, L.L., and R.D. Bradley. 2000. Molecular chondrial cytochrome b sequences. Molecular systematics of the genus Sigmodon. Journal of Phylogenetics and Evolution 16: 366–378. Mammalogy 81: 332–343. Voss, R.S. 1991. An introduction to the Neotrop- Peppers, L.L., D.S. Carroll, and R.D. Bradley. ical muroid rodent genus Zygodontomys. Bulle- 2002. Molecular systematics of the genus tin of the American Museum of Natural History Sigmodon (Rodentia: Muridae): evidence from 210: 1–113. 2009 CARLETON AND ARROYO-CABRALES: ORYZOMYS COUESI COMPLEX 127

Voss, R.S., and L.F. Marcus. 1992. Morphological of the American Museum of Natural History evolution in muroid rodents. II. Craniometric 296: 1–149. factor divergence in seven Neotropical genera, Weksler, M., A.R. Percequillo, and R.S. Voss. with experimental results from Zygodontomys. 2006. Ten new genera of oryzomyine rodents Evolution 46: 1918–1934. (Cricetidae: Sigmodontinae). American Muse- Voss, R.S., L.F. Marcus, and P. Escalante. 1990. um Novitates 3537: 1–29. Morphological evolution in muroid rodents I. Wilson, D.E. 1991. Mammals of the Tres Marı´as Conservative patterns of craniometric covari- Islands. In T.A. Griffiths and D. Klingener ance and their ontogenetic basis in the Neo- (editors), Contributions to mammalogy in honor tropical genus Zygodontomys. Evolution 44: of Karl F. Koopman. Bulletin of the American 1568–1587. Museum of Natural History 206: 214–250. Weksler, M. 2003. Phylogeny of Neotropical Wolfe, J.L. 1982. Oryzomys palustris. Mammalian oryzomyine rodents (Muridae: Sigmodonti- Species 176: 1–5. nae) based on the nuclear IRBP exon. Wolfe, J.L. 1985. Population ecology of the rice rat and Evolution 29: (Oryzomys palustris) in a coastal marsh. Journal 331–349. of Zoology 205A: 235–244. Weksler, M. 2006. Phylogenetic relationships of Zimmerman, E.G. 1970. Karyology, systematics oryzomine rodents (Muroidea: Sigmodonti- and chromosomal evolution in the rodent genus, nae): separate and combined analyses of Sigmodon. Publications of the Museum Michigan morphological and molecular data. Bulletin State University Biological Series 4: 385–454.

APPENDIX 1 Oryzomys couesi aztecus: Mexico, Morelos, Yautepec, 4100 ft (USNM 51171, 51172, 51173 [holotype], 51174, 51175); 10 mi S Cuernavaca (USNM 20907/ SUPPLEMENTARY SPECIMENS EXAMINED 36019). Oryzomyscouesicrinitus:Mexico,DistritoFederal,Tlal- The following specimens were additionally con- pan,7600ft(USNM50181,50182[holotype],50183). sulted apropos evaluation of taxa in western Mexico, Oryzomys couesi regillus: Mexico, Michoaca´n, Los but no decision about their status was reached. The Reyes (USNM 125936–125940, 125943, 125944, pro forma listing as subspecies of O. couesi reflects 125945 [holotype], 125946–125948); 1.5 N Los their current classification (i.e., Hall, 1981), not a Reyes, 4400 ft (KU 96101–96103); 1.5 mi S Los taxonomic recommendation. Reyes, 4500 ft (KU 96104).